Seed banks are facilities that collect, dry, freeze, and store plant seeds so that genetic diversity survives even if the living plants disappear. They range from massive international vaults holding hundreds of thousands of crop varieties to small community collections preserving locally adapted landraces. The idea sounds simple, but keeping seeds alive for decades or centuries involves a surprising amount of biology, logistics, and international coordination, and not every plant cooperates with the process.
The Basic Concept
At its core, a seed bank works by slowing down time. Seeds are living organisms with active metabolisms, and left on a shelf at room temperature, they gradually lose the ability to germinate. By reducing both moisture and temperature, a seed bank pushes that metabolic activity close to zero, extending a seed’s viable lifespan from years to decades or even centuries. Seed banking provides an efficient form of conservation for wild plant genetic diversity, and it serves a parallel purpose for crops by preserving the raw material that plant breeders need to develop new varieties.1Plant Diversity. The contribution of botanic gardens to ex situ conservation through seed banking
There are two broad categories of seed banks. Agricultural genebanks focus on crops and their wild relatives, storing seeds to safeguard food security. Conservation seed banks focus on wild plants, often targeting rare or threatened species. Some facilities, like the Millennium Seed Bank at the Royal Botanic Gardens, Kew, focus primarily on wild species, while others, like the Svalbard Global Seed Vault in Norway, serve as a backup for agricultural genebanks around the world. The distinction matters because the goals, collection strategies, and even the users who request seeds differ considerably between the two.
How Seeds Get From the Field to the Freezer
The journey of a seed into long-term storage follows a fairly standard sequence, though the details vary by facility and species. It begins with collection. For crop genebanks, seeds may come from farmers’ fields, breeding programs, or other genebanks. For wild-plant collections, teams travel to natural habitats to gather seeds from healthy populations, taking care to sample enough individual plants to capture genetic variation without stripping the wild population bare.
Once seeds arrive at the facility, they are cleaned to remove debris, chaff, and damaged material. Then comes the critical step: drying. Seeds are placed in controlled environments where relative humidity is kept low, typically around 15%, which draws moisture out of the seed tissue over days or weeks. The target moisture content for most seeds destined for long-term storage is between 3% and 7% by weight. This level is low enough to halt most metabolic processes without damaging the seed’s internal structures.
After drying, the seeds go into airtight containers, often foil-laminate packets or sealed glass jars, and are placed in freezers. The standard temperature for conventional genebank storage is around minus 20°C, roughly the temperature of a household deep-freeze. At this combination of low moisture and low temperature, seeds that tolerate drying can remain viable for many decades. The proper storage conditions differ for orthodox, intermediate, and recalcitrant seeds, and understanding a species’ tolerance to desiccation and cold is one of the first questions any genebank has to answer before committing to a storage protocol.2PubMed Central. Choosing the Right Path for the Successful Storage of Seeds
Why Some Seeds Cannot Be Banked the Normal Way
The whole system described above works because most crop seeds and many wild-plant seeds are what scientists call “orthodox.” Orthodox seeds naturally dry out at the end of their development on the parent plant, enter a dormant state, and can tolerate being dried further and frozen without losing the ability to germinate later. This is the evolutionary strategy of plants that grow in seasonal climates: produce tough seeds that wait out winter or drought.
But a significant minority of plant species produce “recalcitrant” seeds that cannot tolerate drying. These seeds remain metabolically active and hydrated, and if you force-dry them below a critical moisture threshold, their cells accumulate damage faster than their internal repair systems can handle. In the genus Acer, for example, sycamore produces recalcitrant seeds that lose viability sharply once dried below about 26% moisture content, while the closely related Norway maple produces orthodox seeds with a much more robust internal defense against drying damage.3PubMed. Ascorbate and glutathione metabolism during development and desiccation of orthodox and recalcitrant seeds of the genus Acer Research on cocoa, another species with recalcitrant seeds, suggests the sensitivity to drying is linked to a decline in the enzymes that protect cells from oxidative stress rather than a simple lack of protective sugars.4Seed Science Research. Desiccation sensitivity and activities of free radical-scavenging enzymes in recalcitrant Theobroma cacao seeds
Recalcitrant seeds are disproportionately common among tropical trees. Mango, avocado, lychee, rubber, and many rainforest hardwoods all produce seeds that die if dried. There is also a middle category, “intermediate” seeds, that tolerate some drying but not full conventional bank conditions. Coffee is a well-known intermediate species. For genebanks, recalcitrant and intermediate species represent a genuine gap in coverage. You cannot simply dry and freeze them, so alternative strategies are essential.
Cryopreservation and Other Workarounds
When conventional freezing at minus 20°C is not an option, facilities turn to cryopreservation, which involves storing plant material in liquid nitrogen at minus 196°C. At that temperature, essentially all chemical reactions stop, meaning the material can theoretically be stored indefinitely. Cryopreservation is used within gene banks for the long-term preservation of vegetatively propagated collections, including species that do not produce useful seeds at all, like banana, garlic, and many potato varieties.5PubMed Central. Minimizing the deleterious effects of endophytes in plant shoot tip cryopreservation
For potato, one of the world’s most important crops and one that is propagated clonally rather than from true seed, cryopreservation of shoot tips and pollen in liquid nitrogen has become a standard approach. The most common technique for potato involves vitrification, a process where the liquid inside cells solidifies into a glass-like state without forming ice crystals, which would otherwise rupture cell membranes. Pollen can sometimes be plunged directly into liquid nitrogen without pretreatment.6Plant Breeding and Seed Science. Cryopreservation of Shoot Tips and Pollen of Potato
Cryopreservation is far more labor-intensive and expensive than conventional seed banking. Each sample requires careful preparation, and recovering live plants from cryopreserved tissue demands skilled tissue-culture work. This means cryopreservation tends to be reserved for species where no other option exists, and it handles far fewer accessions per facility than a seed vault can.
Checking Whether Stored Seeds Are Still Alive
Seeds do not last forever, even under ideal conditions. Every genebank runs periodic viability tests, pulling a small subsample of stored seeds and attempting to germinate them under controlled conditions. International standards recommend monitoring germination rates at regular intervals so that when viability starts to decline, staff can grow out a fresh batch of seeds (a process called regeneration) before the stored collection becomes useless.
Getting the timing of these tests right is a balancing act. Test too often and you use up the very seeds you are trying to conserve. Test too rarely and you might discover a collection has already dropped below usable viability. Research at multiple genebanks has focused on making monitoring schedules more efficient by tailoring the intervals to each species’ known longevity under storage conditions, rather than applying a one-size-fits-all calendar.7Plant Genetic Resources: Characterization and Utilization. Safeguarding seed longevity under genebank storage: Evidence based viability monitoring intervals for six tropical species at the Australian Grains Genebank (AGG) Dynamic monitoring intervals reduce unnecessary testing while still catching declining seed lots in time. When a seed lot falls close to or below the threshold for regeneration, curators can also use their professional judgment to schedule shortened follow-up intervals rather than immediately committing to an expensive grow-out cycle.8Genetic Resources and Crop Evolution. A pragmatic protocol for seed viability monitoring in ex situ plant genebanks
The Hidden Risk of Regeneration
Regeneration sounds straightforward: when stored seeds start losing viability, you plant some, let them grow, harvest fresh seeds, and put those back in the freezer. But every regeneration cycle introduces genetic risks. The plants grown out in a field or greenhouse are exposed to natural selection under conditions that may be very different from their original habitat. If only a small number of plants are grown, random genetic drift can shift the genetic makeup of the collection away from the original population. And if certain plants produce more seeds than others, the next generation will be skewed toward those genotypes.
Modeling work has shown that when regeneration is done carefully, with sample sizes of at least 75 plants and bulk harvesting of seed, the accumulation of harmful mutations is minimal. But under less careful conditions, such as growing out fewer than 75 plants or artificially equalizing each plant’s seed contribution, mutation load can increase meaningfully over 25 to 50 regeneration cycles. The situation gets worse if seeds accumulate mutations at elevated rates during long storage itself, in which case even careful regeneration may not fully prevent genetic erosion.9PubMed Central. Deleterious mutation accumulation and the regeneration of genetic resources This is why genebanks try to minimize the number of regeneration events and why maintaining high initial seed viability matters so much.
Wild Plant Conservation and the Millennium Seed Bank
While much of the public attention on seed banks focuses on food crops, conserving wild plants is equally important and arguably harder. Wild species often have poorly understood germination requirements, limited geographic ranges, and small population sizes that make collection risky. The Millennium Seed Bank Partnership, managed by the Royal Botanic Gardens, Kew, is the largest effort focused specifically on wild plant seeds. It conserves propagules primarily from orthodox seed-bearing wild vascular plants from around the world.10Biodiversity and Conservation. The conservation value of germplasm stored at the Millennium Seed Bank, Royal Botanic Gardens, Kew, UK
The practical value of these collections goes beyond simply having seeds on ice. Conserving wild seeds provides insurance for threatened species by enabling their reintroduction into habitats where they have been lost.11Culture, Agriculture, Food and Environment. “The First Step Is to Bring It Into Our Hands:” Wild Seed Conservation, the Stewardship of Species Survival, and Gardening the Anthropocene at the Millennium Seed Bank Partnership If a wildflower disappears from a particular meadow due to development or drought, banked seeds can be the starting point for bringing it back. Wild relatives of crop plants also carry genetic traits, like drought tolerance or pest resistance, that breeders may need in the future but that do not exist in domesticated varieties.
Gaps in Global Coverage
Despite the scale of existing seed banks, coverage of the world’s plant diversity is uneven. Gap analyses in regions like Northeast Africa have found that while many crop wild relatives have some representation in genebanks, depth of coverage is often thin. One study documented that although about half of the taxa examined had at least some accessions in collections, only around 8% were represented by more than 50 accessions, which is generally considered the minimum for capturing meaningful genetic diversity within a species.12Genetic Resources and Crop Evolution. In situ and ex situ conservation gap analyses of crop wild relatives from Northeast Africa In Malawi, the picture was similarly uneven: of 123 taxa of crop wild relatives assessed, only three had collections held at the national genetic resources center.13Genetic Resources and Crop Evolution. In situ and ex situ conservation gap analyses of crop wild relatives from Malawi
These gaps matter because the plants most at risk of disappearing in the wild, especially in tropical and subtropical regions facing rapid land-use change, are often the same plants least represented in seed banks. Filling those gaps requires fieldwork in remote and sometimes politically unstable areas, which is expensive and logistically demanding.
When War Proves the System Works
The most dramatic real-world test of seed-bank infrastructure came from Syria. The International Center for Agricultural Research in the Dry Areas (ICARDA) maintained a globally important genebank in Aleppo holding irreplaceable collections of wheat, barley, lentil, and faba bean landraces adapted to dryland conditions. As the Syrian civil war escalated, ICARDA staff managed to safety-duplicate more than 80% of the collection to the Svalbard Global Seed Vault before the last personnel had to leave the facility in 2014. Beginning in 2015, ICARDA used those Svalbard backups to rebuild their working collections at new sites in Morocco and Lebanon, and they resumed distributing seeds to researchers and breeders internationally.14PubMed. Safeguarding a global seed heritage from Syria to Svalbard
The ICARDA story is often cited as proof of concept for the entire Svalbard model. Svalbard itself does not distribute seeds to users. It functions purely as a backup: genebanks around the world send duplicate samples, and the vault returns them only if the depositing institution loses its primary collection. The vault sits inside a mountain on a Norwegian Arctic island, kept at minus 18°C, with permafrost as an additional safety layer. It holds well over a million seed samples from institutions in virtually every country.
Soil Seed Banks Are a Different Thing Entirely
The phrase “seed bank” also has a completely separate meaning in ecology. A soil seed bank is the natural reservoir of dormant seeds buried in the ground at any given site. These are not collected by humans; they accumulate as plants shed seeds over years and decades, forming a below-ground archive of the plant community’s past and potential future.
Soil seed banks play a crucial role in ecosystem recovery after disturbance. Research on Australian threatened forests found that after wildfire, the soil seed bank may buffer against losses in species diversity even when the standing vegetation is heavily damaged. A significant portion of the species richness in these communities exists solely in the soil seed bank, emerging only when conditions change.15Functional Ecology. The post‐fire recovery of soil seed banks along a fire severity gradient in an Australian threatened mesic forest However, increasing fire severity shifted the composition of both the living vegetation and the soil seed bank, suggesting that repeated extreme fires could eventually overwhelm this natural backup system.
The Native Seed Supply Problem
Having seeds in a bank is only useful if you can get them into the ground when needed. Ecological restoration projects, from mine-site rehabilitation to post-fire revegetation, need large quantities of native seed. And in many countries, the supply chain for native seed is remarkably fragile. A survey of Australia’s native seed sector found that the workforce underpinning it is composed primarily of sole operators and small businesses, and it is significantly underresourced relative to both current restoration needs and likely future demand.16Restoration Ecology. Australian native seed sector characteristics and perceptions indicate low capacity for upscaled ecological restoration
This bottleneck affects seed banks in two ways. First, if restoration practitioners cannot source enough seed commercially, they may turn to conservation seed banks for material that was never intended for large-scale planting. Second, it highlights a gap between the conservation world and the restoration world: having a species safely stored in a vault does not mean you have enough seed to revegetate a landscape. Bridging that gap is an active area of policy debate in countries with ambitious restoration targets.
Who Owns the Seeds
Seed banks hold genetic material from all over the world, and questions of ownership and access are politically charged. The International Treaty on Plant Genetic Resources for Food and Agriculture, often called the Plant Treaty, governs access to a list of major food crops held in international genebanks. Under this treaty, signatory countries agree to share seeds for research and breeding through a standardized material transfer agreement.
For other species, and for commercial uses, the framework is more complicated. The Nagoya Protocol, adopted under the Convention on Biological Diversity, establishes rules for access and benefit sharing. The principle is straightforward: if one country grants access to its genetic resources and another country uses them, the resulting benefits should be shared. The protocol explicitly contemplates commercial benefits arising from the use of genetic resources.17PubMed Central. Access and Benefit Sharing Under the Nagoya Protocol—Quo Vadis? Six Latin American Case Studies Assessing Opportunities and Risk In practice, navigating these rules can be cumbersome enough to discourage researchers from requesting material, which is one of the tensions in the system: access restrictions meant to ensure fairness can inadvertently slow the very research that benefits everyone.
How Genomics Is Changing the Game
A seed sitting in a freezer is only as useful as what you know about it. Historically, genebank accessions came with little more than a label: species name, collection location, maybe some notes on the plant’s appearance. Modern genomic tools are transforming that. By sequencing the DNA of thousands of accessions, researchers can identify which stored seeds carry particular traits of interest without having to grow every one of them out in a field trial.
Work on the USDA Soybean Germplasm Collection illustrates what this looks like. Researchers examined more than 17,000 soybean accessions from China and identified genomic regions involved in environmental adaptation, including genes for flowering regulation, response to day length, and stress tolerance. They then checked whether those potentially valuable genetic variants were present in modern North American and European soybean cultivars. Major genes for early maturity had already been incorporated through breeding, but many other variants associated with adaptation to cold, high-latitude environments were underrepresented in modern lines.18BMC Plant Biology. Using landscape genomics to infer genomic regions involved in environmental adaptation of soybean genebank accessions Findings like these turn a warehouse of frozen seeds into a searchable library of traits, making it far easier for breeders to find the genetic raw material they need as growing conditions change.
Community Seed Banks and Local Knowledge
Not all seed conservation happens in high-tech facilities. Community seed banks, sometimes called seed libraries, are local or regional collections maintained by farmers, indigenous groups, or community organizations. They tend to focus on locally adapted crop varieties, often called landraces, that may not be represented in national or international genebanks. These varieties carry adaptations to specific soils, microclimates, and farming systems that formal breeding programs may overlook.
Community seed banks also serve a social function. They keep seeds and the knowledge needed to grow them circulating within a community. A farmer who loses a planting stock to flood or drought can return to the shared collection rather than buying commercial seed that may not perform well under local conditions. The challenge is sustainability: community seed banks depend on volunteer labor and often lack the climate-controlled infrastructure for true long-term storage. They function more as living, rotating collections than as archival vaults, which makes them complementary to, rather than replacements for, institutional genebanks.
There is growing recognition that the knowledge systems attached to these seeds, the cultural practices around when and how to plant them, which varieties pair well together, how to select seed for the next season, are themselves a form of heritage worth preserving alongside the genetic material. Integrating indigenous and local knowledge with formal conservation efforts remains an evolving challenge, especially when institutional frameworks for data management were not designed with oral or practice-based knowledge in mind.