Seed Moisture Content: definition, measurement and target values

Whether a batch of seed will keep is a question about its moisture content: the share of water in the weight of the seed as it stands. 100 g of seed at 8 % moisture content consist of 8 g water and 92 g dry matter. That figure decides how fast the batch ages. Measured on hemp: at 14 % moisture content and 21 °C no seed germinated after three months, while at 6 % and 4 °C to 10 °C, 81 % to 90 % still germinated after two years.

Fresh weight or dry weight

Two reference bases are in use and they give different numbers. Seed banks work against the fresh weight: fresh weight minus dry weight, divided by fresh weight. Drying technology commonly uses the dry matter as the base, and the same seed then carries a higher figure. 8 % on a fresh-weight basis becomes 8.7 % on a dry-matter basis.

With figures from different sources it pays to check which is meant. In this lexicon the base is always the fresh weight, as is usual in the seed literature.

The measurement destroys the sample

The most accurate method is oven drying. A small sample is weighed, dried out completely and weighed again; the weight lost is the water. The sample is dead afterwards, which is why it stays small: half a gram to a gram, or at least ten seeds, in two separate replicates that may differ by no more than 0.2 %. After the oven the samples cool in a closed vessel with a desiccant so they do not pull moisture back out of the room air.

The temperature follows the make-up of the seed. Oily species are given 103 °C for 17 hours, starchy ones 130 °C to 133 °C for one to four hours. There is a reason for the split: at high temperature oil-rich seeds lose volatile constituents as well, and the measured moisture content comes out too high. Cannabis does not appear in the species list of the standard; assigning it to the gentler method follows from the oil fraction and is an inference rather than a ruling.

This is more than a formality. The three available studies on the moisture content of hemp seed used three different methods: 130 °C for three hours, 103 °C for five hours, and the rule for oily species. By how much their figures diverge as a result has never been determined. Comparing values from two sources may mean comparing two measurement methods.

The practical route: the target weight

Because the oven method consumes the sample, practice works with a calculation. Once the starting value is known, the weight the batch must reach can be worked out: starting weight times (100 minus starting moisture), divided by (100 minus target moisture). From there on it is only weighing.

An example shows which scale serves. 20 g of seed at 8 % are to reach 4.5 %; the target weight is 19.27 g and the difference 0.73 g. That is under four hundredths of the starting weight. A kitchen scale reading in 1 g steps shows none of it, and 0.01 g is the minimum needed. The whole calculation also hangs on the starting value: without one measurement, the weighing runs precisely against an unknown target.

Where such a measurement is out of reach, the detour runs through the air. Instead of determining the moisture content, the humidity that establishes itself in a sealed vessel with the sample is read off. That reading can be repeated as often as wanted and destroys nothing.

Which values apply

For long-term storage the seed banks name 3 % to 7 % depending on species, with oily species at the lower end, and 3 % to 8 % for collections drawn on regularly. For hemp there is a series over two years: 6 % and 8 % at 4 °C to 10 °C held 81 % to 90 % germination, while 14 % at 21 °C led to total loss within three months.

Downwards there is a limit below which further drying gains nothing and can even speed up ageing. It depends on the species and sits low in oil-rich ones: 3.3 % in soya, 2.7 % in flax, 2.4 % in niger, 2.0 % in groundnut and sunflower, against 4.5 % to 5.0 % in rice, wheat and barley. For cannabis it has not been determined.

What rapid moisture meters deliver

Devices that display a moisture content within seconds do not measure it; they infer it from an electrical property. They have to be calibrated against the oven method for each species, and they are more reliable between 6 % and 25 % moisture content than below. Long-term storage turns on the range under 6 %, and that is where they are least dependable.