Harvesting, Cleaning and Storing Seed – the sequence in six steps

Six working steps lead from the ripe plant to the freezer. Each ends at a condition you can see or weigh: the seeds drop out of the flowers, the green is gone, the weight holds still, the empty seeds have been carried off, the target weight is reached, the vessel is shut.

Drying happens in stages, and that sets the order. First the flower material dries until it releases the seeds. Between the two sieving passes the husks still clinging to the seeds dry off. Then the seed comes into equilibrium with the room air, which sharpens the density separation and supplies the starting value for the weighing. Last it goes to the level it will be stored at.

Two percentages run alongside each other here. Moisture content says how much water sits inside the seed; the balance shows it. Relative humidity says how damp the air beside it is; the hygrometer shows that. The hygrometer is needed once, in the drying room.

Harvesting and drying until the seeds drop

Ripe seeds are hard and have turned from green to brown. The period given is about 30 to 35 days after flowering, from a speed-breeding trial, and 27 to 46 days after full flower for field stands.

Two routes for drying. Either hang the whole plant upside down, or strip the flowers from the stems and lay them out separately. The second is faster and separates the plant mass from the seed earlier. The international genebank standard asks for freshly harvested seed to go into controlled drying within three to five days and not to be exposed to more than 30 °C or 85 % humidity until then.

The moment has come when the seeds drop out of the flowers under light pressure and rubbing. The small stems snap when bent. If nothing falls out and the material still bends, it is too early; rubbing now crushes seeds and only breaks up the green.

Sieving in two passes

Work with a set of screens, here 5.0 mm, 4.5 mm, 3.6 mm, 3.0 mm, 2.5 mm and 1.8 mm. Two of them sit on top of each other: you rub through the upper one, and whatever passes collects on the lower.

First pass. Rub the dry flowers over the coarse screen under light pressure until everything has gone through. On the screen below, the seeds stay behind, along with seeds that still carry their husk. Only enough pressure for the seeds to come free: the coat must be neither scratched nor crushed, because damaged seeds lose viability quickly.

Interim drying. Put what has been sieved into shallow trays, about 60 cm by 40 cm at 7 cm deep, and leave it for a few days. In that time the clinging husks dry off and come loose.

Second pass. Rub again, now over the finer stages. What has come loose falls one stage lower; the remains of the husks pass off as grit through the finest mesh. Between 2.5 mm and 3.0 mm this set is missing a stage that would hold back the poorly developed seeds. Which meshes suit depends on the variety; measured seed dimensions run 4 mm to 5 mm long by 3 mm to 4 mm across, with a reported range across varieties of 2 mm to 6 mm by 2 mm to 4 mm.

Hand winnowing. In the same tray, push the seed into one corner, hold the tray at an angle and blow across it. The light material lifts off, the seeds stay put. That clears the loose green.

Drying to equilibrium and weighing

Spread the seed thin, or put it in an open, rolled-down bag that lets air reach the seeds. Weigh at the same time every day. While the figure falls, the exchange is still running. Once it holds steady over several days, the end point of that air is reached. A single quiet day is not enough: the laboratory only counts equilibrium once three successive weighings lie within 0.1 percentage points of each other, because water moves out of the seed only slowly.

Now read off the humidity and the temperature of the room. From those the moisture content follows without sacrificing a sample in an oven, because at equilibrium humidity and temperature fix it. At 51 % humidity and 26 °C that comes to a calculated 8.0 %.

Those 8.0 % are calculated and not measured. Two points have been measured on hemp, at 34 % and at 94 % humidity, and 51 % sits between them. At the lower point, at 35 °C, the calculation matches one of two varieties exactly and sits 0.3 percentage points under the other.

From here the balance leads.

Sorting by size and density separation

Before the density separation, sort by size once more, so that only density varies in the air stream.

In the aspirator the seed falls against a rising air stream. Gravity pulls down and depends on weight, the air stream pushes up and depends on size. Among seeds of equal size, weight and therefore density decides, and that is what the machine separates by: empty seeds the same size as full ones are carried off, while two seeds of equal density cannot be told apart. A screen cannot do this, because it sorts by measurement, and a small full seed lands in the same fraction there as a large empty one.

The previous drying step pays off here: a moist seed is heavier than the same seed dry, and while a batch has dried unevenly the density clouds overlap. That follows from how the machine works and has not been measured.

Adjustment runs through the bypass air inlet: if everything comes out at the bottom the suction is too weak, if everything ends up in the collector it is too strong. For a batch that cannot be replaced, run two passes, the first with weak suction so no good seed goes with it, the second sharper. Weigh both fractions: that fixes the reject share as a figure.

Fine drying to the target weight

Now it goes below what room air provides. For hemp the measurement series give a target range of 5 % to 7 % moisture content, centred on the 6 % at which two storage series found their gain; drying further brought no additional advantage there. The same range holds for the freezer, and there rather not its lower end, because colder storage raises the most favourable moisture content. A rule of thumb from seed storage sets the lower limit: between 4 % and 14 %, every percentage point of water removed doubles seed life, and below 4 % drying can do harm in some species. How those figures come about, and which humidity leads to which moisture content, is covered in the article on drying seed.

The target weight is fixed as soon as the starting value is known:

target weight = starting weight × (100 − starting moisture) ÷ (100 − target moisture)

100 g at 8.0 % gives a target weight of 97.9 g for 6.0 %, so 2.1 g have to go; for 5.5 % it is 97.4 g and 2.6 g. Those amounts are small, which is why 0.01 g resolution is needed. A balance in 1 g steps shows the same figure from day one.

Four routes lead there, and they differ in how far they can go.

A dried room. A dehumidified room dries the seed with nothing in the vessel and nothing to watch, down to the value its air sets. Genebanks keep such rooms at 15 % to 20 % humidity; at room temperature hemp would come out there at a calculated 4.2 % to 4.8 %, below the target range. For hemp, 25 % to 30 % is enough and gives 5.5 % to 6.1 %. The route is still more expensive than keeping pre-dried seed in tightly sealed containers.

Silica gel. Its capacity depends on the humidity: at 25 % a kilogram takes up 125 g of water, at 45 % already 260 g. For drying, add it at the same weight as the seed; that amount brings a starting value of about 16 % down to just under 6 %. For merely keeping seed dry in a sealed vessel a tenth of the seed weight suffices, and even that pulls a batch at 6 % below 5 % initially. Too much of it overdries, and the target weight is the brake.

The beads go into a cloth bag rather than loose among the seeds, because mixing produces abrasion. That matters most for the blue grade: its colour indicator is cobalt chloride, and the genebank manual warns of its possible carcinogenic effect and advises handling it under an extraction hood where dust is likely. As alternatives it names orange beads that turn colourless, or low-dust bead gel; whether the dye of the orange grades is harmless has not been checked here. Undyed silica gel carries no dye. Regenerate it away from the seed in an oven; for silica gel 100 °C to 175 °C are given.

Anhydrous calcium chloride. Works like silica gel and costs little. After use it is discarded or reused to make a saturated salt solution. Also into a bag or its own dish, not into the seed.

A saturated salt solution. It stands in an open dish inside the closed container and sets a fixed humidity there: lithium chloride holds 13 % at 20 °C, calcium bromide 18 %, calcium chloride 30 % at 25 °C. Because it sets a humidity rather than drawing water, it never dries below the value that humidity dictates, and the batch needs no watching. Which salt to use follows from that value: calcium chloride brings hemp to a calculated 6 % or so, in the middle of the target range, lithium chloride to 3.9 % and calcium bromide to 4.6 %, below it. The seed must not touch the solution.

Drying calculator

Once the weight holds steady over several days, the seed stands in equilibrium with the room air. Humidity and temperature then give its moisture content, and from that the weight it still has to lose to reach a target.

What was measured

What was measured

Read the humidity in the drying room, not inside the vessel. A hygrometer in a closed jar shows a different value as soon as a desiccant sits in there.

Where it stands now

Where it should end up

Target moisture

The target range for hemp is 5 % to 7 %, derived from two storage series that found their gain at around 6 %. It holds for fridge and freezer alike; for the freezer the evidence leans against the lower end. Below 4 % drying can do harm.

All target levels

The row for the selected target moisture is highlighted.

Target moistureTarget weightStill to lose

The moisture content is calculated from an estimating equation and not measured; the calculator assumes an oil fraction of 32 %. Two points have been measured on hemp: at 34 % humidity and 35 °C the calculation matches one of two varieties, at 94 % it runs almost three percentage points high. The target weight itself is plain arithmetic and as accurate as the balance.

Sealing and freezing

Whether the value reached will hold depends on the container: on its material and on its closure. Which seven are in common use, where each one’s weak point sits and which three mistakes cost more than the choice of material is covered in the article on seed storage containers.

For small individual lots a separate rule applies: a desiccant there is not meant as an addition. A commercial sachet weighs a multiple of twenty seeds and goes on drying the batch rather than merely showing its state; a tenth of the seed weight already pulls a batch from 6 % below 5 %. Anyone wanting to keep an eye on the moisture over the years puts the indicator into a larger outer container holding the individual packs, not into each one.

Then cool it down and freeze it. When a container later comes out of the freezer, it is opened only once it has reached room temperature, otherwise water from the room air condenses on the cold seed.

Anyone wanting to test the batch should wait or pre-chill. In two varieties, fresh hemp seed carried a dormancy of about 35 to 42 days after harvest: 21 days after harvest 79 % and 91 % germinated, although a staining test showed 97 % and 96 % living seeds. Five days at 10 °C on moist paper before the germination test lifted the figures to the level of the staining test, and from about four weeks after harvest pre-chilling made no further difference.

What has not been established

For cannabis no published screen specification exists; the set above comes from the practice of a single operation and needs checking against the variety. A comparison of hand and machine threshing on hemp is missing, and the rule against crushing comes from general genebank practice. That pre-drying sharpens the density separation follows from how the machine works and was not measured. And below 34 % humidity no measurement of the equilibrium exists for hemp, so the conversion from humidity into moisture content is calculated there.

Sources