A Guide to Efficient Long Term Seed Preservation

Types of Seeds

From the point of view of their preservation, two types of seeds are usually distinguished: orthodox and recalcitrant. Orthodox seeds can be desiccated down to moisture contents of 4-7% and even ultradesiccated down to contents of 1-3%. In this way, the moisture factor can be used without restrictions to achieve good longevity. Recalcitrant seeds do not tolerate desiccation well, thus making other procedures necessary. However, no clear limits between both types of seeds exist and there are cases showing an intermediate behavior.

General Principles

Low moisture, low temperature, low ethylene concentration, and probably low oxygen concentration are the most relevant factors to have in mind for efficient long-term seed preservation. Some interrelations among them exist.

Low moisture and low temperature have been the traditional factors considered. Still, in the practice of the past few decades, the low temperature has comparatively received much more attention, while efficient control of low moisture has been largely neglected. However, in order to guarantee efficient long-term preservation of orthodox seeds, good control of seed moisture content is essential. In this respect, two major principles should be taken into account:

The container should be perfectly tight, preventing water vapor intake. If the container is not perfectly vapor-tight, the seeds will tend to get balanced with the external air humidity. In this way, any potential benefit of low temperature will be offset by an increase in seed moisture. It must be noticed that relative humidity inside any uncontrolled cold room is usually very high. Further to this, it should be emphasized that dried seeds are highly hygroscopic, and – even if often very slowly – there is time enough ahead for moisture to enter the container. Unfortunately, the utilization of inadequate containers has been commonplace in seed genebanks (Gómez-Campo, 2002).

Ultra drying down to a moisture content of 1-3% might extend the seed life span perhaps 4-16 times (Harrington, 1972). Ultra drying has been considered harmful by some authors (Walter & Engels, 1998; Hu, C. & al. 1998), while others have defended that it is not damaging to orthodox seeds (Ellis, 1998; Hong & al. 2005). The debate might last for years until the subtle factors behind these opposite opinions are fully understood. However, under the conditions utilized in the Universidad Politécnica de Madrid (UPM) genebank, ultradry seeds preserved for 40 years have maintained their viability intact (average = 98,4%, not significantly different from 100%; Pérez-García & al., 2007). As far as we know, these results are unmatched by any other seed genebank and set a solid base to develop improved preservation methods. This has moved the author to prepare the present condensed guide.

When dealing with ultradry seeds, enormous savings of energy can be made by avoiding the use of too low temperatures during storage. The temperature may be the key factor when seeds are merely dried but it loses relative importance for ultra-dried seeds (Pérez-García & al., 2007). Independently, very low temperatures, close to that of liquid nitrogen (-196 ºC), provide an alternative method for the preservation of seeds and tissues (cryopreservation) that might play an important role in the future, especially for recalcitrant material.

A third factor – the presence or absence of oxygen – has been the source of contradictory literature over the years and, perhaps, for this reason, it has largely been neglected from the practical point of view. However, recent research by Ellis & Hong (2007) suggests that ultradry seeds may be sensitive to oxygen; therefore, it is a factor that, at least tentatively, should be taken into account. Perhaps, the success of the UPM bank partly resides in the fact that air was originally substituted by CO2 in the atmosphere within the containers.

Ethylene and other toxic gases that slowly evolve during seed aging should be removed in order to extend the seed life span.

Choice of containers

For many years it has been widely admitted that containers that are appropriate for cold drinks or to store chips for a few months are also appropriate for long-term seed storage. This is very far from reality. Results of a survey on 40 different containers – some of them widely used in many genebanks – showed that 36 of them (90%) allowed moisture inside in less than three years (Gómez-Campo, 2002, 2006a).

All plastic containers allowed moisture inside. The water molecule is small enough to get through the pores of the polymers used to make the plastics and in the medium and long term, the fact is unavoidable – although not noticeable in the short term. Polyvinyl bags are especially permeable to humidity. An additional reason for the failure of containers with a lid is based on the fact that both pieces are usually made of different materials, therefore expanding or contracting in a different way with changes in temperature. This creates fissures allowing humidity to get inside the container. This last mechanism appears not only in plastic containers but also in most glass containers with twist-off or screw lids. Even plasticized twist-off type jam jars often show oxidation stains inside after some time has elapsed.

A special mention should be made to aluminum foil bags because of their still widespread use in genebanks of cultivated species. It’s been a long time since bi-laminated bags (coated with plastic only on one side) were rejected on the basis of their inefficiency. However, tri-laminated bags (coated on both sides) are still widely accepted. According to our experience, a maximum of 80% of a given set keeps vapor tight after 3-6 years. It is evident that to base the whole operation of a genebank on a material with these characteristics is risky. To put an example forwards, a collection of 100.000 samples where we learn that 20% of them (20.000) won’t keep their tightness creates a really awful scenario which is now quite common in genebanks. Additionally, it is not possible to see from the outside which ones the problem is. Massive germination tests or even measurements of moisture in all the samples are not practical solutions. Let’s add to it that, being sealed with plastic, they carry on the long-term disadvantages of this material.

The aforementioned experiment showed that only four containers remained vapor-tight (Figure 1):

containersa. Flame-sealed glass vials. Used in the UPM’s genebank since 1966, they have only very exceptionally (0,3%) shown cases where moisture had come inside, due to the sealed tip having got broken. This tip can be protected with a mixture of resin and wax, as will later be explained. Undoubtedly this is the best option for small valuable accessions of wild or crop species such as rare or endangered species or genotypes, crosses, mutants, and so on.

b. “Kilner” (“Scotch”) jars with a rubber joint and propped-up lid, whose use is quite spread in kitchens. After 12 years, only one out of 25 units allowed moisture inside. Its size allows their use with samples of every size, including those bulky ones often kept in genebanks of cultivated species. A good adjustment of the joint should be ensured and also that nothing sticks between the glass and the joint (adhesive tape, little seeds, or gel grains). It is possible that some joints need to be replaced after several years. As they are hermetic, seed samples can be stored inside in bulk, or in any type of smaller container such as paper or polyvinyl bags.

c. Some glass flasks with a plastic lid, usually used in laboratories to store chemicals behaved very well during the experiment but other very similar ones behaved very badly. The difference was due to the quality of the plastic joint under the lid. Their opening being narrow, these jars are somehow uncomfortable to be used with medium-sized or large seeds.

d. Heat-sealed, metal cans coated with plastic on the inside. This inside plastic coating may be beneficial to prevent any possible damaging effects from the metal vapors in contact with the seeds over the years. The seal is also plastic. On the other hand, these containers are opaque and do not allow external control with colored moisture indicators – unlike the three previous ones.

Those vials with a rubber lid fastened by an aluminum cover used some years ago to contain penicillin, were not tested. However, users of this container refer to its behavior positively.

It is very possible that other appropriate containers could be found in the future. However, to do reliably so, perhaps several years might be necessary. On the other hand, models very similar in appearance – case (C) and also some (B) – can behave very differently and it is advisable to test them before choosing a particular one. The method there described might serve to test other containers.

Considering that most banks around the world have too often used non-vapor-proof containers, this experiment probably shows the main reason for failure in seed preservation – especially for banks of cultivated species. It also shows some possible solutions. Solution (B), in particular, provides a larger volume and opens the way to the use of the silica gel method in seed genebanks of cultivated species – a method already been used successfully since 1966 with wild species (Gómez-Campo, 1972).

On the other hand, it is obvious that the container’s tightness is linked, in a practical way, to the need for frequent or only occasional handling of the seed lots. For very long-term collections of valuable materials addressed to the future (“black-box” collections), there is no doubt that the flame-sealed glass vials (Figures 1A and 2a) are the best option. Those can also be of help for other materials used more often, especially from wild species or any small-sized species, by placing smaller quantities of seeds and making more vials, so that one of them can be broken when necessary.


For base (long-term) collections, Kilner jars (Figures 1 B and 3) are also advisable and so are the other options already mentioned. Active collections (for distribution), require more frequent access and need containers capable of being opened and closed. Kilner jars may also be used here if being fast when taking the sub-samples and, whenever possible, doing it in a dry environment.

For active collections of species with small seeds, tubes with a screwed plastic lid (Figure 2b) might be of interest. They are easy to use but not entirely reliable.

Transparency of the containers becomes a key quality when wishing to take advantage of the benefits of the silica gel method for easy and direct monitoring of the samples from the outside.

Ultra drying

To this date, the most commonly used procedure has consisted of drying the seeds with different procedures until reaching a moisture content of approximately 5-7%. After this, the seeds are placed in containers whose vapor tightness is most often untested and the success is mainly trusted to low temperatures. Ultra drying has been very rarely used, perhaps by no more than 50 banks – mostly of wild species in botanical gardens – out of more than 1 500 existing in the world. The UPM’s results have proved beyond doubt the efficiency of ultra-drying in orthodox seeds at least under anaerobic conditions and also prove the comparatively lesser importance of temperature for ultra-dried seeds.

It should be noted that a moisture content of 4-5% may be enough for the efficient conservation of some orthodox seeds such as those of many legumes (Ellis & al. 1988). However, as this author points out, these seeds do not suffer when ultradesiccated to 1-3%. In a gene bank, with thousands of accessions being handled, the need to individualize the methods should be avoided. In this light, a general benefit to the use of ultra drying can be attributed – always speaking of orthodox seeds.

Ultra drying with silica gel

Ultra drying could be achieved in different ways, but perhaps the most practical one might consist of storing the seeds in balance with dehydrated silica gel. Leaving some gel afterward with the stored seeds inside the tight container is most useful because the colored indicator would show any accidental entry of moisture.

Table 1. Advantages of the use of silica gel

  1. It provides a practical method to desiccate the seed samples.
  2. It reaches moisture levels of 2-3%, lower than those obtained by most.
  3. It keeps these levels indefinitely in tight containers.
  4. It warns about possible anomalies (i.e. moisture intake) in the container – when used in combination with a colored indicator.
  5. It can be regenerated after its use through a new dehydration process with heat.
  6. It delays also aging by absorbing toxic gases produced during the process.

Table 1 offers a summary of the advantages of using silica gel with an indicator. Besides those related to ultra drying itself, it is worth noting the last one (number 6), which refers to a totally independent protection mechanism. In seed samples stored within a closed container, damaging gases are generated during the aging process and as they accumulate, they prove deathly for the entire sample. The mortality, which might have been supposed a priori probabilistic for every individual seed, becomes accelerated at the end. Being capable of absorbing these gases, silica gel significantly delays aging and postpones the moment of death. Ethylene is an important component of this mixture of gases (Lee, & al. 2001).

Silica gel is simply silicic anhydride, SiO2, though amorphous (non-crystalline) obtained through an industrial process. It is granular in texture, white, and very porous. It is this last characteristic that gives its absorbent properties. When it is well dehydrated and placed on a scale, it can be observed to absorb up to 20% of its own weight of water. If placed in a closed container such as a Kilner jar, it balances itself with the confined atmosphere until this reaches approximately 10-12% of relative humidity.

The name “gel” is here improperly used. We understand by gel some substance of a colloidal nature, which is not the case. However, the use of this term for this product has long been used and is now strongly established.

silica gel

The color usually shown by silica gel is due to an indicator, added to see directly when it is dehydrated and when it has absorbed moisture. For many years, cobalt chloride (Cl2Co) has been used. This substance gives the dehydrated gel a strong blue color and a pale pink color to the gel that has absorbed moisture. Recently, the European Union banned its use because of considering it carcinogen through inhalation. A search for new alternatives led to some iron salts, where the change in color can be poorly distinguished. At present, the most advisable alternative is methyl violet, which gives the dehydrated gel an orange color and a green color to the hydrated gel.

The gel is manufactured in granules of different sizes. This site is not irrelevant because, to be used in Kilner jars, a gel with a size > 4 mm can be more convenient, while a 2-3 mm size can be better for glass vials. By hydrating with water coarse silica gel granules (be careful, heat is produced!) the grains get fragmented, and it is possible to screen them in fractions with smaller granule sizes.

Hydrated silica gel can be regenerated through heating. When it is accompanied by cobalt or iron salts, this is done at approximately 220 0C, thus a standard cooking oven can be enough. The process may last one or two hours – depending on the mass to be regenerated. With an organic indicator (e.g. methyl violet) a temperature of 230-248°F must not be surpassed and the process will require comparatively more time.

Freeze drying

Freeze drying (“lyophilization”) has been successfully used to obtain ultradry levels of moisture. This procedure consists in freezing the seed water and sublimating the ice afterward. This technique was developed in the National Centre of Natural History (Paris) in the 1980s for long-term storage of pollen grains (Schoenke & Bey, 1981. The Conservatoire Botanique National Méditerranéen (CBNM) started using it for seeds in Porquerolles in 1987 (Figure 6).

It is assumed that freeze-dried seeds do not suffer any decrease in their viability because of the process. However, it does not seem to be the case for every species, and the details when applying this technique must be optimized for every species (Anonymous, 2006). As it will be explained, freeze-drying may have the advantage of achieving a direct ultra drying instead of one done in stages – as is generally the case with silica gel which needs to be regenerated every time it is used.

In comparison, freeze drying needs a rather expensive machine (Hu, X. & al. 1998). In turn, silica gel reaches lower water content. Also, considering that silica gel is going to play an essential role afterward in keeping low moisture and monitoring the samples, it may be more practical a direct drying with gel. In other words, while freeze drying only ultradries, silica gel not only ultradries but also provides efficient monitoring during subsequent storage.

The low-temperature factor

The use of low temperatures has been widespread in long-term seed preservation. Standards set by the Food and Agriculture Organization of the United Nations (FAO) and the International Plant Genetic Resources Institute (IPGRI, now Biodiversity) advise a storage temperature of -64°F or lower (FAO/IPGRI, 1994) and almost everybody now planning a genebank instinctively thinks of a cold room capable of at least reaching -68°F. Localization of the new installations of the Nordic Seed Bank in a region of “permafrost” responds to the same idea.

Unfortunately, emphasis on low temperatures and neglecting low moisture (either because of not drying enough or because of the use of inappropriate containers) has brought that a significant number of the actual genebanks are worried about the burdensome need to rejuvenate their prematurely aged material.

Seeds that had been freeze-dried in the Conservatoire Botanique National Mediterranéen (CBNM) of Porquerolles and then kept at room temperature for 10 years showed slightly better behavior than those kept under low temperature. In a much longer term, ultradry seeds with silica gel kept for 40 years at room temperature inside a closet (Pérez-García & al., 2007) did not behave very differently from those kept in the cold room. It appears hence clear that the main role attributed to low temperature during decades should now be replaced by a much more thorough attention to the moisture factor.

However, it is obvious that low temperature helps, and it would be irresponsible not to take it into consideration. In the last mentioned work, it is suggested that temperatures between -23°F and +41°F could very well be enough for orthodox ultradry seeds, thus saving considerable amounts of energy.

The practice of long-term seed preservation

An alternative sometimes used consists of controlling moisture in the cold room itself. It is not a bad idea itself, although it shows some disadvantages. With a temperature of about 32°F (more than sufficient as it has just been explained), reaching a relative humidity of approximately 25% inside the chamber is neither difficult nor expensive. Anyway, this would be far more efficient than having a chamber at -68 /-77°F with the tightness of the container being neglected. However, controlling the moisture content within the containers themselves by using silica gel is, in the long term, much more practical and efficient, since it allows reaching more suitable ultradry levels and better monitoring.

Ultra drying with silica gel can be achieved by putting the samples inside a closed container along with the dehydrated silica gel (Figures 2 and 3).

The use of flame-sealed vials, obtained from standard laboratory glass tubes of 20 ml is a system whose high efficiency has already been proven. If it cannot be generally recommended, is only on the grounds that it allows little volume – a maximum of 8 ml – for large accessions. However, it should be considered that 8 ml means some 160 seeds of wheat, 1 400 of cabbage, and 40 000 of Diplotaxis dietitian (extinct in Nature). The method is almost perfect for especially valuable material such as rare, endemic, or threatened species, special breeding material, “black-box” collections, etc. Also, the idea of preparing several vials for each accession is realistic and most often worthwhile.

In the following paragraphs, the different steps to create such ampoules according to the procedure used in the UPM genebank (Figure 5) are explained in detail:

  1. The test tubes must be made of alkaline glass. This can be asserted through some greenish color in their rim. Other glass types might soften more slowly or can even crack when heated.
  2. Since long-term preservation is intended, it is wise to place it inside an internal label with the accession number, because any external label could get deleted or lost after a time. The number is easily written with a ballpoint pen on the sticky side of a label and then pressed against the wall with a rod.
  3. The seeds, already clean and dry, will be introduced from another tube or by means of a funnel. Although it is possible to fill up to 8 ml, a less quantity is often placed in order to have more vials available when it becomes necessary to break one in the future. Rubber stoppers in Figure 5 are intended to avoid provisionally moisture intake from the outside.
  4. A piece of cotton, preferably hydrophobic and previously dehydrated will be used to separate the seeds from the silica gel which will be placed on top.
  5. Afterward, approximately 3 ml of dehydrated silica gel will be added. Before moving on to the next step, it is advisable to place a rubber cap on the tube and wait for 15-20 days. If the seed is not properly dry, this will be shown in the bottom of the gel and simply changing this gel could be enough to take the sample to the desired moisture balance.
  6. Another piece of cotton, like the one used in (4), will help to keep the whole in place. Around 5 cm of the tube should be left free, to allow sealing it by heat.
  7. Numerous groups of tubes can then be placed (convenient) for 24 hours within a box that is filled with CO2. Any other inert gas could do the job but CO2 being heavier than air may be managed more easily. Cylinders with dry CO2 are easily obtained from the market.
  8. The flame from a standard laboratory burner may not be enough for a comfortable sealing of the tube. Therefore, a second entrance hole will be useful to apply a stream of oxygen or air. The current supplied by a standard house hairdryer could be enough to stoke up the flame.
  9. The sealing itself can be done reasonably quickly once some practice has been acquired. Holding the tube by its base with the fingers (no heat is felt) and with a long tong on the upper rim, concentrating the flame action and turning to spread the heat around, approximately 100 vials per hour can be thus closed.
  10. Immersion of the sealed vials in water for 24 hours, would allow checking if any of them was not properly closed – the silica gel will change color. The procedure should then be repeated for this particular case.
  11. It would only remain necessary to add an external label with more data, according to the preferences of each bank, and to protect somehow the more delicate part of this set, the flame-sealed tip. Both goals can be combined. A long, sticky label is rolled around the top and is used as a plank mold to pour inside a solidifying liquid.
  12. Some substances were tried, such as melted wax, epoxy resins, etc. The conclusion reached was that the most practical solution consists of a mixture of wax (2/3) and pitch (= colophony or resin) (1/3) – this last substance used to harden the wax. The mixture is melted and poured with a dropper in the hole created by the label around the glass tip.
  13. In order to avoid the top surface becoming scratched, a nail varnish could be applied.

When stretched with the flame and externally labeled, the vials should become of final homogeneous size. In the UPM’s genebank, they have traditionally been stored inside Kilner jars – long before finding that the last also were tight enough. Around 25 vials can be placed in each jar – thus obtaining double security.

The described protocol might seem difficult and work demanding but this is not at all the case when the work is properly set up and organized. The seeds can thus be preserved without any evident viability loss for at least 40 years, which may well mean efficient preservation for at least one century or probably more. The advantage and the disadvantage at the same time are that the vials can only be opened by breaking them, unlike what happens with the Kilner jars.

The detection of sufficiently reliable larger containers as the Kilner jars (Figure 3) opened the way for the adaptation of the silica gel method to the larger samples that are common in crop genebanks. In this case, three important differential aspects should be taken into account

Firstly, drying in situ is trickier than it was with tubes. Since the seed sample may have an initial moisture content of approximately 8-12%), the gel will be hydrated by the sample and will change its color. It will be necessary to replace this gel with a new, dehydrated one, and the operation might need to be repeated several times. Once the gel does not change color, one more replacement is convenient. The seed will then be ultradesiccated and will remain indefinitely balanced with the dry gel if the container is kept tight. However, changing the gel repeatedly might be fastidious. Therefore, it is convenient to dry the seeds as much as possible by some other method (sun exposure, dry room, dry air currents, or other drying agents such as Cl2Ca). In other words, it is advisable to start ultra-drying with the seeds already as dry as possible. Drying speed will depend on the mass ratio between the desiccating agent and the seeds but once the equilibrium is reached, only a small amount of silica gel to act as an indicator is enough (and also convenient, see later).


Plastic chambers (not long-term tight, but tight enough to perform this operation, see figure 5, can hold a series of cardboard boxes, some with gel and others with seeds. By changing the gel when necessary, it will be possible to desiccate inside a high number of samples in only a few weeks. Since most seed-collecting missions at least in the Mediterranean area take place in June or July, it is possible to use the period of August – September to desiccate the seeds.

If there are funds available to purchase new equipment, and once the efficiency of the method becomes well established, a previous freeze drying (Figure 6) can take the seeds directly to moisture levels very close to the balance with dehydrated gel.

Ready-to-use desiccators containing silica gel which is automatically regenerated already exist in the market. They usually consider that the chamber where the samples would be placed will be provided by the user (Figure 7). When the chamber is a store room, even tiled, big or small, it is inevitable that humidity will come inside through the walls. This makes it difficult to achieve relative humidity levels below 15%, especially with low temperatures. Besides, the drier constitutes a dynamic system with several pieces, joints, etc., very far from the simplicity of the jar of Figure 3 (where <5% can be reached). Compact systems with built-in chambers to be placed on top of a laboratory bench are already present in the market. However, these cabinets never guarantee a relative humidity of under 20%. R.H. under 15% could perhaps be reached in the future with well-designed impermeable glass or metal chambers. However, if the final goal is to store the seeds in Kilner (Scotch) jars, previous storage at 15-20% R.H. in the cold room may help to reduce the number of times the silica gel must be replaced.

Secondly, it is important to take into account that silica gel can absorb important amounts of water (up to 20% of its own weight) and lose part of its efficiency before changes in color can be appreciated. This is particularly important with jars that are frequently opened and closed. To obviate this problem fully dehydrated silica gel should always be used and, once the equilibrium seed/gel has been obtained, to place only a small amount of gel (i.e. a small bag in the upper part) is sufficient. In this way, any moisture intake will be detected soon.

In the third place, it is convenient to minimize the presence of oxygen, either by filling the containers as much as possible with the seed material or by replacing the air with an inert gas, preferably CO2 because it is heavier and tends to occupy the lower parts of a volume as water do. A few open Kilner jars could be placed in an “ad hoc” box where CO2 is added. Otherwise, CO2 could be directly added from the cylinder if this is accompanied by a pressure-reducing device. Also, unlike sealed vials, a Kilner jar may be opened, perhaps frequently, and this has to be taken into account by adding CO2 once in a while. It should be noted that the substitution does not need to be perfect because the aim is to reduce (not necessarily eliminate) the O2 present within the container.

The same system (Kilner jars with silica gel at the bottom) could be valid to stop aging in badly preserved seed collections. It would be enough to place the samples inside (in bulk, in their original containers, or better in paper envelopes) and to regenerate the gel as many times as necessary until no further color change would occur. Additional substitution of the gel will ensure that this is completely dehydrated. In the particular case of foil bags, several of them could be fitted inside each jar. Each jar would be afterward given a number and the accessions position within the genebank then recorded in the database.

Recalcitrant seeds

These seeds are characterized by their relatively high moisture content (15-25%) and by a clear intolerance to desiccation. (Berjak & Pammenter, 2004). Therefore, they cannot be preserved by desiccation.

Recalcitrant seeds can often be recognized morphologically because of their large size and smooth tegument. However, other seeds with a more normal appearance could also be recalcitrant. In the Mediterranean area they are not very spread but they can be found in woody species, climactic or not (Quercus, Castanea) or in riverbank species (Populus, Salix). In the rainforest, the species with recalcitrant seeds are dominant. Ecologically, they follow a strategy of lasting as seedlings, better than as seeds. That means that their seeds avoid getting too dry in a humid environment, where the very process of drying would be more difficult. Their lifespan is generally short, between a few weeks and 2-3 years.

If we make an exception for those species growing in riverbanks, seed size seems to be a key factor to determine the orthodox or recalcitrant behavior of a species. This is because mechanical damage produced during drying is more severe in large seeds. For example, the seeds of pumpkin (Cucurbita sp.) and of some Phaseolus species (P. lunatus) can be damaged through drying while the smaller watermelon seeds (Citrullus moschatus) or those from green beans (Phaseolus vulgaris) are orthodox. The internal composition could also set differences. When working with small seeds, it is possible to be fairly confident that they will be orthodox, whereas, with an increase in size, exploratory tests should be carried out.

Being orthodox or recalcitrant is not an absolute character (Berjak & Pammenter,1994), since “semi-recalcitrant” seeds with an intermediate behavior exist. These accept a partial desiccation, if only to a certain level. In any case, they should not be expected to tolerate ultra-drying.

To preserve recalcitrant seeds, there is not much to be done, except to put them in their lower tolerance limit as far as moisture content and temperature are concerned. As they inhabit wet places, it is usual that they die prematurely due to fungal or bacterial attacks. This means that treatment with an appropriate phytotherapeutic product could perhaps extend their life span to some extent. Perhaps the most practical method whenever possible could be the establishment of gene banks of seedlings since these seedlings use to be quite long-lived when placed in low-lighting conditions. For long-term effects, it would be necessary to resort to the cryopreservation of their embryos.

It is not difficult to establish a protocol to set apart orthodox from recalcitrant seeds – based on the first ones being tolerant to desiccation. This could consist of the following steps: (1) An initial germination test, (2) the desiccation of an equivalent lot, and (3) a new germination test of the desiccated seeds to show whether they are still alive or not. However, it should always be kept in mind that low values in germination tests do not necessarily mean that the sample is dead or dying. These values could be originated from dormancy. If so, immediate treatment of the seed (either by scarification, immersion in gibberellic acid solution, or any other suitable procedure) should follow in order to overcome any possible dormancy. This is particularly important for wild species.

Before deciding whether a given seed is recalcitrant – especially in medium-sized seeds – there is still a test to conduct, which may enable us to desiccate it and preserve it in the long term. Some apparently recalcitrant seeds seem to tolerate drying and perhaps ultra-drying when these are gradually carried out, without sudden changes. For these “pseudo-recalcitrant” seeds it is advisable a previous successive balance with atmospheres showing a decreasing relative humidity. If this occurs, the number of species that could be preserved in the long term by the silica gel method will considerably be enlarged.


It consists of storing the material at temperatures near that of liquid nitrogen (-320.8°F). Under these conditions, all enzymatic processes are practically halted and it is thought that any type of biological material (in plants: meristems, calluses, embryos, pollen, seeds, somatic tissues, etc.) can thus be preserved for a potentially infinite period. Only cosmic rays could place a limit after a few millennia.

Orthodox seeds can easily be cryopreserved. However, doing this has not had much meaning once the high efficiency of ultra drying has been demonstrated for them. Ultra drying is also much cheaper. It is with recalcitrant seeds that cryopreservation might play a role in the future.

For efficient cryopreservation, it is fundamental to avoid the intracellular formation of ice crystals which are highly damaging to the cell’s internal structures. A rapid descent of temperature is a must in the process, and a previous desiccation or ultradesiccation of the material is highly helpful. Also, cryoprotective treatments (osmotic agents and/or substances such as glycerol, proline, abscisic acid, etc.) are often used. The most commonly used methods today aim to obtain a transition between an aqueous solution and an amorphous “glassy” state by means of concentrated solutions of sucrose and some derivatives of glycerol.

In the case of recalcitrant seeds, many difficulties appear with whole seeds (Engelmann, 1997) but these can be partly overcome by preserving the embryos only (Fu & al., 1993). Basically, the technique consists of an encapsulation of the embryo in calcium alginate followed by a desiccation in concentrated sucrose solutions and then a final immersion in liquid nitrogen.

Additional comments

Ultra drying and storage with silica gel of orthodox seeds in hermetic containers, preferably in an anaerobic atmosphere, with only moderately low temperatures, provide a tested and reliable method to keep the seeds alive for many decades and – likely – for centuries.

Efficient long-term seed preservation saves much genetic material, and also much time, labor, electricity, and money. The now widespread resignation over the need to rejuvenate the seed material in cycles of 25-35 years (in fact middle term preservation), could give a path to the quite different situation of long-term cycles of one or more centuries where the tremendous inconveniences of rejuvenation become minimized. Also, much of the time devoted to too frequent and tedious germination tests – so often only valid to check passively how the stored seeds are aging and dying – can be devoted to research on more useful and stimulating subjects, such as the initial quality of the seed accessions, the possible dormancy in some of them and the adequate method to remove it, the orthodox, recalcitrant or pseudo-recalcitrant behavior of others and so on (Gómez-Campo, 2006b).


This guide is an English-free translation of the revised chapter devoted to seed preservation which will be included in the Spanish version of Manuale APAT, Roma. Thanks are due to M. Pilar Ballesteros who made valuable style suggestions.


  1. Anonymous (2006) Manuale per la racolta, studio, conservazione e gestione ex situ del germoplasma. Agenzia per la protezione dell’ambiente e per i servici technici (APAT), Roma.
  2. Berjak, P. & Pammenter, N.W. (1994). Recalcitrance is not an all-or-nothing situation. Seed Science Research 4, 263-264.
  3. Berjak, P. & Pammenter, N.W. (2004). Recalcitrant seeds. Benech-Arnold, R.L. & Sánchez, R.A. (eds.) Handbook on Seed Physiology. Haworth Press. Binghampton, N.Y. pp. 305-345.
  4. Ellis, R.H. (1998). Longevity of seeds stored hermetically at low moisture contents. Seed Science Research 8 (supplement No. 1), 9-10.
  5. Ellis, R.H., Hong, T.D. (2007). Seed longevity – moisture content relationships in hermetic and open storage. Seed Science and Technology 35, 423-431.
  6. Ellis, R.H., Hong, T.D. & Roberts E.H. (1988) A low moisture content limit to logarithmic relations between seed moisture content and longevity. Annals of Botany 61, 405-408.
  7. Engelmann, F. (1997) Importance of desiccation for the cryopreservation of recalcitrant seed and vegetatively propagated plants. Plant Genetic Resources Newsletter 112, 9-18.
  8. FAO/IPGRI (1994) Genebank standards, Rome. Food and Agriculture Organization of the United Nations/International Plant Genetic Resources Institute.
  9. Fu, J.R., Xia, Q.H. & Tang, L.F. (1993) Effects of desiccation on the excised embryonic axis of three recalcitrant seeds and studies on cryopreservation. Seed Science and Technology 21, 85-95.
  10. Gómez-Campo, C. (1972) Preservation of West Mediterranean members of the cruciferous tribe Brassiceae. Biological Conservation 4, 355 360.
  11. Gómez-Campo, C. (2002) Long-term seed preservation: the risk of using inadequate containers is very high. Monographs ETSIA, Universidad Politécnica de Madrid 163, 1-10.
  12. Gómez-Campo, C. (2006a) Erosion of genetic resources within seed genebanks: the role of seed containers. Seed Science Research 16, 291-294.
  13. Gómez-Campo, C. (2006b) Long-term seed preservation: updated standards are urgent. Monographs ETSIA, Universidad Politécnica de Madrid 168, 1-4.
  14. Harrington, J.F. (1972) Seed storage and longevity. Seed Biology. Vol. 3, pp. 145-245 in Kozlowsli, T.T., New York, Academic Press.
  15. Hong, T.D., Ellis, R.H., Astley, D., Pinnegar, A.E., Groot, S.P.C. & Kraak, H.I. (2005) Survival and vigor of ultra-dry seeds after ten years of hermetic storage. Seed Science and Technology 33, 449-460.
  16. Hu, C., Zhang, Y., Tao, M. & Chen, S. (1998) The effect of low water content on seed longevity. Seed Science Research 8 (supplement 1), 35-39.
  17. Hu, X., Zhang, Y., Hu, C., Tao M. & Chen, S. (1998) A comparison of methods for drying seeds: vacuum freeze-drier versus silica gel. Seed Science Research 8 (supplement 1), 29-33.
  18. Lee, P.-C., Taylor, A.G., Zhang, M. & Esashi, Y. (2001) Evolution of volatiles during seed aging: exogenous gas application. Journal of New Seeds 2, 77-91.
  19. Pérez-García, F., González-Benito, M.E. & Gómez-Campo, C. (2007) High viability recorded in ultra-dry seeds of 37 species of Brassicaceae after almost 40 years of storage. Seed Sci. and Technology 35, 143-153.
  20. Schoenike, R.E. & Bey, C.F. (1981). Conserving genes through pollen storage. Franklin, E.C. (ed.) (1981) “Pollen management handbook”. Forest Service Agriculture Handbook, USDA, 287, 72-73.
  21. Walters, C. & Engels, J. (1998) The effects of storing seeds under extremely dry conditions. Seed Science Research 8 (supplement 1), 3-8.