Drying Seed – Target Value, Duration and Control
Seed, meaning seeds kept back for sowing later, ages faster the more water it holds, at least across the usual storage range. Drying seed therefore means taking enough water out before storage to bring it to 5 % to 7 % of its weight, roughly a quarter of what freshly harvested hemp seed carries. How dry it gets is set by the air in the room, when it has got there is shown by a scale, and the method stays the same whether the seed then goes into the fridge, the freezer or a sealed jar in the cupboard.
A two-year storage series on one hemp variety shows how much the water matters, with both batches kept at 21 °C, ordinary room temperature. Seed at 14 % water no longer germinated at all after three months. The same variety dried to 6 % still germinated at 89 % after 24 months. The temperature was the same for both; the difference was the water content.
What moisture content room air settles at
Moisture content is the share of a seed’s weight that is water. Seed lying open in a room gives water off to the air until the two are in equilibrium. The moisture content it settles at is called the equilibrium moisture content, and the same air draws no further water out beyond it. Which value that is depends on the relative humidity: how much water vapour the air carries compared with the most it could hold at its temperature. At 20 °C the calculation gives these values for hemp:
| Humidity in the room | Moisture in the seed | Relative to the 5 % to 7 % target |
|---|---|---|
| 15 % | 4.2 % | below |
| 20 % | 4.9 % | just below |
| 25 % | 5.5 % | on target |
| 30 % | 6.1 % | on target |
| 40 % | 7.2 % | just above |
| 50 % | 8.2 % | above |
| 60 % | 9.4 % | too wet |
| 70 % | 10.6 % | too wet |
| 80 % | 12.0 % | too wet |
These figures are calculated, not measured. An estimating equation from a genebank manual derives them from humidity, temperature and the oil fraction of the seed, here taken as 32 %; measured across 20 hemp genotypes, the oil fraction ran from 28.5 % to 36.0 % with a mean of 31.9 %. A curve that plots moisture content against humidity like this is called a sorption isotherm. For hemp it has only been measured between 34 % and 94 % humidity: at 35 °C and 34 % the variety Altair sat at 5.8 % moisture, a second variety 0.3 to 0.5 percentage points higher, and at 10 °C and 94 % Altair sat at 13.7 %. The equation matches the first point at 5.8 % and overshoots the second clearly at 16.5 %. Below 34 % humidity no hemp measurement exists, and that is exactly where the dry rows of the table sit.
The gaps between rows widen further down. From 30 % to 50 % humidity the moisture content climbs by 2.1 percentage points, from 60 % to 80 % by 2.6. Humid air pushes the result up faster than dry air brings it down.
In practice: a room at 25 % to 30 % humidity hits the target range, at 40 % the seed stays just above it, and from 50 % it stays above it however long it sits. A hygrometer, a small meter for air humidity, shows what a room holds. If the reading is too high, three routes remain: a drier room, a dehumidifier in the room, or a closed container in which a desiccant such as silica gel takes up the water or a saturated salt solution holds the humidity at a fixed low value. How the container methods work in detail is covered in the article on harvesting, cleaning and storing seed.
What humidity an indoor room holds depends on whether cold outside air comes in and is warmed. Cold air carries little water vapour; brought up to room temperature its relative humidity drops although not a drop has gone anywhere. A heated room in the cold season is therefore drier than the same room unheated. Which months those are depends on the hemisphere; the mechanism holds everywhere.
How long drying takes
Freshly harvested hemp seed carries 22 % to 30 % water, three to six times the target range. 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. None of the work read here offers a formula that predicts drying time in room air from humidity and layer depth. What exists are two measured points, and they start from different values.
The first comes from the laboratory, where seed is dried in a thin layer with warmed air and a fan. Hemp seed first wetted to 20 % water lost 0.2 to 0.3 kg of water per kg of dry matter per hour in the fastest phase at 40 °C and 30 % humidity, dry matter being the seed without its water, and came down to 4.5 %; the longest runs took over ten hours. Below 35 °C the rate stayed under 0.2 kg per kg per hour in every run; warmer air dries faster, so a tray at room temperature takes longer. The second point is closer to everyday conditions: hemp seed at 7.7 % and 10.0 % water, spread 5 mm thin, reached about 5 % after two weeks at 22 °C to 28 °C and 30 % to 45 % humidity.
Drying slows towards the end. Below 13 % moisture content the rate in the laboratory trial fell under 0.1 kg per kg per hour, less than half its peak. The last percentage points therefore take more time than the first.
Two things stretch this out. The measured values come from thin layers; in a deep pile the air reaches the middle less well, and how much longer that makes drying has not been measured in the work read here. And where the starting moisture is above 15 %, as in freshly harvested seed, the genebank manual calls for two-stage drying: slowly at somewhat higher humidity first, then the actual drying. In soya, groundnut and chickpea, drying too quickly has cracked the seed coat. No such cracking has been documented in cannabis.
How to tell when drying is finished
Since the duration cannot be worked out in advance, the end point is measured, and there are four ways to do it.
The simplest: weigh the tray of seed at the same time every day. As long as the weight falls, the seed is still giving off water. Once it stays the same over several days, the seed has reached the end point of that air, its equilibrium moisture content, and waiting longer changes nothing. A single quiet day is not enough: water moves slowly to the surface of the seed, the laboratory only counts equilibrium once three successive weighings lie within 0.1 percentage points of each other, and fast thin-layer drying ended on average 0.5 percentage points above the true end point. This needs no known starting value, but it does need a scale reading to 0.01 g: 20 g of seed going from 8 % to 4.5 % moisture lose a calculated 0.73 g, under four hundredths of their weight, and a scale in 1 g steps cannot follow that.
Anyone after a particular target value works out a target weight: starting weight times (100 minus starting moisture), divided by (100 minus target moisture). An example: 100 g of seed at 9 % water that should reach 6 % are on target at 100 × 91 ÷ 94 = 96.8 g. The same formula set the moisture levels in the two-year hemp series. It assumes the starting moisture is known, and the oven test supplies that: two small samples of 0.5 g to 1.0 g each, or at least ten seeds, are weighed, dried out completely and weighed again after cooling; the weight they lose is their water. For oily species the rules prescribe 103 °C for 17 hours, against 130 °C to 133 °C for non-oily ones, because hotter ovens drive volatile compounds out of oil-rich seed and make the moisture content look too high. Cannabis is not on the species list of those rules; placing it with the oilseeds follows from its oil content. The samples do not survive the oven, and the rest of the batch never goes in.
The third way measures the air above the seed instead of the seed. A small digital hygrometer sealed in a jar with a handful of seeds shows, once its reading stops changing, which humidity the seed stands in equilibrium with; the table above converts that into moisture content. The seed comes to no harm, and the reading can be repeated later on the stored jar. How accurate consumer meters are below 20 % humidity has not been tested in the work read here.
The fourth way saves measuring on later batches. Anyone who has once recorded how a batch dries under given conditions can read the duration off that curve for further batches of the same species under the same conditions.
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
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
All target levels
The row for the selected target moisture is highlighted.
| Target moisture | Target weight | Still 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.
What moisture content to aim for
Two storage series on hemp set the frame. One ran 24 months on a single variety: at 6 % and 8 % water, germination capacity, the share of seeds that still germinate, stayed largely stable over the whole period at 4 °C to 21 °C, ending at 81 % to 90 % at 4 °C to 10 °C. The other ran 66 months on five forms of cannabis at 11 %, 6 % and 4 % water and four temperatures from 20 °C down to −80 °C. According to its published summary, both drying to 6 % and cooling to 5 °C brought a large gain, drying further to 4 % brought none on top, and cooling further a small one. A rule of thumb from seed storage sets the lower limit: between 4 % and 14 %, every percentage point of water removed doubles seed life, below 4 % drying can do harm in some species, and below about 2 % drying damage becomes likely in many. Together that gives a target range of 5 % to 7 %: centred on the 6 % at which both series found their gain, with a margin to the 4 % limit. This is derived from those statements, not measured in its own right; in none of the hemp series read here did a value below 6 % show any advantage.
Storage temperature does not push the target downwards. The colder the storage, the less the moisture content matters: at −20 °C it affected germination less in the two-year series than at the warmer levels. The international genebank standard also states that a lower storage temperature raises the most favourable moisture content and increases the risk of overdrying. For the freezer, the evidence therefore leans against the lower end of the range. The worry that seeds could be damaged by ice in the freezer concerns seed at around 15 % water and more; seed dried to 4 % to 8 % sits well below that.
Seed banks, which keep seed for decades, follow an international standard: dry to equilibrium at 5 °C to 20 °C and 10 % to 25 % humidity, then store at −18 °C. The standard sets the air rather than the moisture content, because the same air leaves less water in oil-rich seed such as hemp than in starchy seed such as wheat. Applied to hemp, its range therefore works out at a calculated 3.4 % to 6.2 % moisture content, and its dry half lies below the 5 % to 7 % target range: 10 % humidity at 5 °C leads to 3.9 %, 25 % humidity to 5.5 % to 6.2 %. Anyone following the standard for hemp therefore takes its humid end, around 25 %. None of this has been measured on hemp, because the standard’s entire humidity range lies below 34 %, the driest point measured on hemp.
How far overdrying does harm has not been established for hemp. A laboratory method that ages seed artificially at 65 °C placed the point below which further drying stops helping low in oil-rich species: 3.3 % for soya and 2.0 % for groundnut, against 4.5 % to 5.0 % for wheat and barley. The manual that gives these figures itself warns against carrying them over to real storage conditions unchecked.
What comes after drying
Whether the value reached will hold is decided by the packaging. Many materials let water vapour through, and a batch brought to 6 % drifts back towards the room humidity over time in the wrong packaging; which material does what is covered in the article on seed storage containers. When a batch comes out of the fridge or the freezer, the container is opened only once it has reached room temperature. Otherwise water from the warm room air condenses on the cold seed, the way it does on a bottle taken out of the fridge.
What has not been established for cannabis
Four gaps belong with the figures above. Below 34 % humidity there is no measurement of equilibrium moisture content for hemp; the dry rows of the table are calculated there. No model in the work read here predicts drying time in room air from humidity and layer depth. Whether drying too fast cracks the seed coat of cannabis, as in other oilseeds, has not been tested. And the three hemp studies read here determine moisture content with three different oven methods, so their figures may be offset against one another by an unknown amount.
A further open question is whether the standard’s cool drying is the best choice for freshly harvested seed. In soya beans a first drying stage of up to eight days at 40 °C, 20 °C above the standard’s upper limit, extended later longevity in an accelerated ageing test in three of four seed origins, in one of them from 29.7 to 66.6 days. In cowpea the advantage appeared in only three of ten batches. Soya holds about 21 % oil, cowpea about 1 %; hemp at 28.5 % to 36.0 % sits closer to soya. In rice the effect appeared in seed that still held more than about 16.5 % water at harvest, and freshly harvested hemp seed at 22 % to 30 % lies above that. On cannabis this is untested, and until then cool drying remains the supported recommendation.
Sources
- Tiwari and Jian 2023: Desorption and Sorption Isotherms of Different Varieties of Hemp Seeds, Agriculture 13, 1959
- Cockson and colleagues 2025: Impact of seed moisture and temperature on hemp seed germination, Agrosystems, Geosciences and Environment 8
- Rao, Hanson, Dulloo, Ghosh, Nowell and Larinde 2006: Manual of Seed Handling in Genebanks, Bioversity International, Handbook No. 8
- FAO 2014: Genebank Standards for Plant Genetic Resources for Food and Agriculture, revised edition, Rome
- Willan 1985: A Guide to Forest Seed Handling, Chapter 7 Seed Storage, FAO Forestry Paper 20/2
- Whitehouse and colleagues 2018: Further Evidence That the Genebank Standards for Drying Orthodox Seeds May Not Be Optimal for Subsequent Seed Longevity, Biopreservation and Biobanking 16
- Galasso and colleagues 2016: Variability in Seed Traits in a Collection of Cannabis sativa L. Genotypes, Frontiers in Plant Science 7, 688
- Small and Brookes 2012: Temperature and Moisture Content for Storage Maintenance of Germination Capacity of Seeds of Cannabis sativa, Journal of Natural Fibers 9(4)