Growing Garlic from True Seed
Ted Jordan Meredith and Avram Drucker
First published in the Seed Savers Exchange Spring 2012 edition of
The Heritage Farm Companion
First published in the Seed Savers Exchange Spring 2012 edition of
Garlic growers sometimes refer to garlic cloves that are reserved for planting as “garlic seed,” but what we want to talk about here is garlic from true seed---the product of sexual reproduction. Garlic seed is a bit smaller than onion seed, but otherwise resembles it. In the first generations of garlic seed production, growing garlic from seed is not particularly easy, but neither is it out of the scope of the average grower---and with subsequent generations of seed-produced garlic, the process becomes much easier, as we will discuss later.
Why even bother growing garlic from seed when growing garlic from cloves is much easier? Asexual reproduction, growing garlic from cloves or bulbils, produces a genetically identical clone of the mother plant. This can be desirable for ensuring uniform continuance of a preferred cultivar. However, if asexual reproduction becomes the exclusive method of reproduction, as has essentially been the case with garlic over the centuries, the implications become quite negative.
As with any plant species that is repeatedly reproduced primarily by asexual means over an extended period, genetic diversity and adaptation is severely constrained and the species is ultimately at risk. Sporadic asexual mutations may enhance diversity to a limited degree, but nothing of the order possible through sexual reproduction. Existing planting stock also frequently suffers from a buildup of pests and diseases. Most garlic cultivars, for example, carry some degree of virus infection that is passed on during asexual reproduction. Garlic from seed-produced plants can reduce or eliminate diseases and pests and significantly increase plant vigor and yield.
Mature garlic umbel, ovary, and seeds.
Garlic was long thought to be sterile, but in 1875 Eduard Regel described unique flowering characteristics in garlic found in the wild, raising the prospect that some garlic might still be capable of producing true seed. As early as the 1950s, research groups within the former Soviet Union reported limited success in producing a few seeds from garlic strains originating in Central Asia. By the 1980s, limited production of garlic seed was reported in Japan, Germany, and the United States. The first garlic seed experiments were highly labor intensive and yielded very few seeds. In recent times, with refined methodologies, literally millions of viable garlic seeds have been produced by researchers and agribusiness interests, largely for experimental purposes.
If agribusiness companies choose to make seed available at some point, it is highly likely that it will be hybrid seed rather than open-pollinated seed in order to protect profits and the large investment involved in bringing the seeds to market. Some new garlic strains have been patented, but little has seen the light of day as new cultivars or seed available to the average grower.
We can change that.
Garlic reproduces itself in three ways, bulb division, bulbils from topsets, and seed. It is likely that some garlic in the wild still reproduces via seed as well as asexually, but it appears that even in the wild garlic plants strongly favor asexual mechanisms. Garlic collected in the wild by plant explorers and researchers generally need human manipulation to successfully produce seed. Not only in cultivation, but also in the wild, it appears that garlic’s capacity for sexually reproducing itself may be diminishing. Interestingly, and fortunately, this propensity is reversible. Subsequent generations of garlic produced via sexual reproduction begin to shift reproductive tendencies back toward sexual reproduction via seed.
Although the scientific community and agribusiness have produced garlic seed somewhat routinely, it appears that very few garlic growers and Seed Savers Exchange members are active in the process. Agribusiness has its own goals, such as virus free stock, patented cultivars, and F1 hybrids. Preserving and enhancing genetic diversity may not necessarily be an agribusiness imperative---but it is clearly an imperative for Seed Savers Exchange members. For growers, the ability to produce a crop that is totally free of viruses and various pests such as nematodes is another major benefit of seed produced lines of garlic.
Producing garlic from seed can be challenging, particularly if certain elements are missing from the process. We offer the following methods not because we are experts, which we surely are not, nor because these are ultimately the best methods, which they likely are not. We offer these methodologies because we know that they produce results, and because we want to encourage others to produce garlic from seed and to discover and share even better methods.
Although many have contributed in many ways, we would specifically like to acknowledge the assistance we have received from Dr. Philipp Simon, United States Department of Agriculture (USDA) research leader and professor of horticulture at the University of Wisconsin; Dr. Maria Jenderek, Plant Physiologist, USDA Agriculture Research Service, Fort Collins, Colorado; and Barbara Hellier, Horticulture Crops Curator, USDA Agriculture Research Service, Pullman, Washington. Without their advice and counsel, we would not have succeeded in producing garlic from seed. We owe them our profound thanks.
Spathe removed, showing an umbel tightly packed with bulbils and flowers. The bulbils in this Marbled cultivar are relatively large and easy to remove.
Growing garlic from seed has essentially two major phases---producing the seed and growing the seed out into fully-developed plants that are capable of reproducing via cloves or bulbils, or again via seed. Both phases present challenges, but the challenges are hardly insurmountable.
Which garlic cultivars are capable of producing seed? Non-bolting or “softneck” cultivars such as those in the Artichoke horticultural group are clearly not candidates, since by definition they do not produce the necessary flowering structures. It had once been thought that only a relatively few bolting cultivars had the ability to produce seed and that most of those would fall into what we would regard as the Purple Stripe horticultural group. Genetic studies have shown that the Purple Stripe group comprises the most ancestral garlic forms currently in existence. They are closest to the Central Asian origins of the species, and they are the antecedents of all the cultivars we know today. However, with increasing understanding of garlic seed production we now know that many cultivars in several horticultural groups are capable of producing seed.
Of the generally recognized horticultural groups, the Marbled Purple Stripe group has been the most reliably productive for us. This may be because Marbled garlics generally have a thicker scape (flower stalk) than Purple Stripe garlics. Most of our seed production is done with severed scapes kept in water rather than by leaving the whole plant in the ground (more about this later). Severed scapes must be able to sustain umbel (flower head) viability. This may favor the more robust scapes of Marbled garlics. Those who leave the entire plant in the ground during the seed production process may well find the productivity balance shifting toward the Purple Stripe garlics.
Cultivars from other horticultural groups are far less reliable and productive, but we have produced seed from cultivars in the Porcelain group. We have even coaxed a few seeds from cultivars in the Glazed and Rocambole groups, but these latter two groups are quite marginal at best. The following is an alphabetical list of some of the cultivars in general circulation that have been known to produce seed. The numbers in parenthesis are USDA accession identifiers: Bai Pai Suan, Brown Rose, Brown Tempest, Brown Vesper, Chimyon (Z 059), Darcheli, Duganskij, Floha, Krasnodar White, Metechi, Mexican Red (the cultivar with this name that generally resembles a Porcelain garlic, aka Azataza), Novotroitsk (W6 26171), Punuk (W6 27971), Rosewood (PI 493099), Shatili, Shvelisi (aka Chesnok Red), Sural (W6 26172), Tien Shan (PI 615416), Yampolskij (PI 540340), and Verchnjaja Mcara (PI 540356, aka Red Czar). This list should be regarded as merely illustrative rather than exclusive. Many other cultivars in general circulation are likely capable of producing seed, particularly those in the Purple Stripe and Marbled groups. Garlic cultivars having flowers with purple anthers are much more likely to be male fertile and to produce seed than those with yellow anthers.
We have had the most prolific results from several of the more recent Central Asian accessions to the USDA garlic germplasm collection. These generally have not yet been classified or else have characteristics that do not readily place them in the existing horticultural groups. They include Sural (W6 26172), Chimyon (Z 059), and Punuk (W6 27971). Of the more commonly available cultivars, Brown Vesper has been particularly productive for us.
Most garlic growers remove the scapes of bolting garlic to ensure the plant’s energy is directed to the bulb for maximum size. For garlic intended for seed production, the scapes and the developing umbels must be retained. Garlic growers stop fertilization when bulb development is underway and stop or reduce watering as the plant leaves senesce and harvest nears. For garlic intended for seed production, senescence must be delayed as long as possible and the plants should be given ample water. Continued fertilization may also be helpful.
As a bolting garlic plant matures, the scape emerges and first coils and then uncoils as its umbel develops. Garlic umbels have both bulbils for asexual reproduction and flowers for sexual reproduction. Bulbils look like tiny garlic bulbs or cloves. Bulbils and flowers compete for the plant’s resources. With certain exceptions, if the plant is left on its own the bulbils win and the flowers wither and die before they can produce seed. The bulbils must be removed from the umbel in order to tilt the balance toward seed production. Removing the bulbils allows the flowers to develop, achieve anthesis, and produce seed. Interestingly, in subsequent generations of seed-produced plants the bulbils are often far fewer and may not require removal for successful seed production.
|Tweezing out the bulbils so that only flowers remain.|
One can quickly become a connoisseur of tweezers for bulbil removal, finding a favored design just as one might settle upon a favored garden spade or rake. Tweezing out the first bulbils is the most difficult. Begin with a bulbil that protrudes somewhat prominently and try to penetrate down to grasp it at its base or along its full length and pluck it out. A few flowers are usually lost in the initial process, but as an area is plucked free of bulbils it provides a space for nearby bulbils to be rocked toward the open space to dislodge them. Bulbil removal is a combination of plucking them out with tweezers and rocking them out to dislodge them. Continue working the umbel until all of the accessible bulbils are removed. If an area is particularly problematic, it can be left for a day or two until the bulbils become less compacted and more accessible. Flowers near remaining bulbils will tend to wither more rapidly. Because bulbils as well as flowers are usually still developing at the time bulbil removal begins, it is usually necessary to return a week or so after the initial procedure to remove any bulbils that subsequently developed.
Bulbils removed, leaving only the flowers.
Most fertile garlic cultivars have purple anthers.
Garlic can be left in the ground for the entire seed production process, or the scapes can be severed from the plant just above the last leaf blade and kept in a container of water. Whether to leave the entire plant in the ground or sever the scape is a matter of preference and environmental conditions. Although it was once thought that leaving the scape attached to the plant caused the developing umbel to suffer in competition with the garlic bulb for the plant’s resources, this does not seem to be an overriding concern. Additionally, a severed scape bears a major burden of sustaining the umbel through the lengthy process of fertilization and seed maturity without the benefit of a supporting root system.
All things being equal, there is some evidence that leaving the scapes unsevered best sustains the umbel for seed production. All things are seldom equal, however, and various factors enter consideration. Stooping over garlic plants in the field in the heat of summer for the tedious process of bulbil removal is certainly one consideration. Other environmental conditions affect the plant itself. Heat sensitivity varies among cultivars, but during anthesis, the period when the flowers become sexually viable, temperatures exceeding approximately 85°F over a period of days can whither flowers and reduce pollen viability. Conversely, temperatures below 65°F can reduce or prevent seed set. A daily period of temperatures between 70°F and 75°F likely approximates an ideal range for optimal results. Mites, thrips, wind, and rain can also damage the delicate flowering structures and inhibit pollination. Mitigating mechanisms can be employed to optimize conditions, such as shade cloth to reduce heat, sunny protected areas to increase warmth or reduce wind, and the like. Some of the mitigating mechanisms are most easily provided if the scapes are severed and kept in transportable containers of water. Although we generally sever the scapes, we have had good results with both methods.
The severed scapes with bulbils removed are kept in water.
Fertilizing the water the scapes are placed in or spraying the scapes with liquid fertilizer is sometimes recommended, though we have had good results without. Severed scapes may dry and lose color or become slimy toward their base during the lengthy seed maturation process. If this occurs, simply trim away the brownish or slimy portion of the scape, shortening as necessary so that a viable part of the scape is in water and can continue to sustain the umbel.
Garlic flowers are protandrous--- individual flowers are incapable of fertilizing themselves. The anthers release pollen two to four days before the stigma on the same flower becomes receptive. However, the full inflorescence of a garlic plant achieves anthesis over a period ranging from some five to twenty-five days. Thus, even though individual flowers are incapable of fertilizing themselves, the flowers of a single garlic plant’s inflorescence can fertilize other flowers in the inflorescence. Garlic flowers are typically pollinated by insects. Anything from honeybees to houseflies can accomplish the task. If necessary, in an isolated environment, a small paintbrush and tedious repetition will also work.
Fertilization complete, the swollen ovaries will yield seed at maturity.
Each garlic flower ovary has three chambers. Each chamber contains two ovules, thus a maximum of six seeds per flower is possible. The ovaries swell as the seeds develop, looking much like the seed ovaries of onions or other alliums. Seeds are ready to harvest approximately 45 to 60 days after pollination. From a practical standpoint, we harvest the seeds when the umbel and seed ovaries have completely dried, usually in October or November. In other climates, drying may occur more quickly. Initial seed yields are typically quite low, but subsequent generations of seed-produced plants yield significantly more seeds, sometimes more than 600 per umbel. Early efforts may be difficult, but later efforts can be exceptionally rewarding.
Growing seed into plants
Seed from garlic plants that have previously been propagated only by asexual means (cloves, bulbils) have a low germination rate ranging from 10% to an optimistic 35% at best. In our own efforts, we have achieved approximately 13% germination with first generation seed. First generation plants tend to produce a higher frequency of small seeds that are frequently inviable. Viability is relatively poor even for larger seeds. However, studies have shown that subsequent generations of seed-produced plants typically have a much higher germination rate, sometimes as high as 100%.
First generation seedlings often exhibit a high frequency of unfavorable characteristics, such as stunted growth, deformed leaves, limited root development, and chlorophyll deficiencies. Subsequent generations of seed-produced garlic exhibit increasing vigor and a lower frequency of genetically deficient plants. This is not to say that the first generation efforts are doomed to weakling garlic. We have certainly seen a number of weak first generation plants, but we have also seen exceptionally vigorous plants as well. For example, a seed planted in late winter of 2010 was harvested as a large round (a bulb that is undivided into separate cloves) in August of that year. The round was planted again in the fall and harvested as a fully-developed bulb in the summer of 2011. The plant was exceptionally vigorous and appeared to be virus free. Its bulb measured 2½ inches in diameter.
New garlic seedlings in germination tray.
Garlic seed has a period of dormancy and should not be planted immediately after harvest. This is likely an adaptive response to garlic’s native origins. If seeds sprouted prior to the onset of a Central Asian winter, their chances of survival would be greatly reduced. Exposure to a period of cold shortens dormancy. An artificial period of moist cold in a refrigerator sufficiently mimics a natural period of cold exposure in Central Asia.
Garlic seeds should be given a bleach soak prior to planting to help protect them from contamination, followed by a cold treatment to shorten dormancy. Soak the garlic seeds in a 1% solution of household bleach (1 teaspoon bleach in 2 cups water) for 20 minutes, rinse the seeds, distribute the seeds on moist paper towels, place the seeds in a plastic sack, and store in a refrigerator for approximately four weeks. We informally tested the importance of this step by planting 26 seeds without the bleach and cold treatment. Only one seed weakly germinated and subsequently died. At the same time, we planted 51 seeds that had received the bleach and cold treatment. The germination rate of those seeds was 13.7%, about what was expected for first generation seeds.
Seedlings transferred to small pots.
After the refrigerator cold treatment, we plant the seeds in trays under artificial lights and cover the trays with a clear plastic dome. Ambient air temperature is approximately 65°F and no supplemental heat is employed. It is likely that the seed will successfully germinate in a reasonably wide temperature range, but we have no additional experience in that regard. We have never had a problem using good quality potting soil, but a sterile seed starting mix is probably a safer option. We cover the seed with about ¼ inch of soil. The first seedlings emerge in about 6 days. We have had weak seedlings emerge as long as 2 months after planting. Most of the viable seeds will emerge within about 2 weeks. Expect some of the first generation seeds to die because of genetic flaws, such as deficient chlorophyll. The range of seedling vigor is substantial. Some will likely show exceptional vigor.
As soon as conditions permit, plant the seedlings outside in the ground. Garlic seedlings are quite hardy, but there are limits. Like all plants, garlic seedlings should be hardened-off before transitioning to the outdoors. In colder or harsher climates a greenhouse or cold frame may be requisite intermediate steps. Depending on outdoor growing conditions, moving the new plants directly from seed tray to the outdoor garden bed may be possible, but transplanting the seedlings into small pots is often a beneficial or necessary intermediate step. Always make sure that the seedlings have ample room for unimpeded root growth.
A garlic plant started from seed earlier the same year.
The penny at the base shows the plant’s size.
Ideally the new garlic plants will grow vigorously and be ready for harvest in summer or early fall, senescing their leaves from the lowest leaf upward, just like a garlic plant grown from a clove. We have experienced some variability in this regard, particularly with late-planted seeds. Some plants may continue to resemble a green onion late into fall without senescing or developing a round. This is not a desired outcome, but these plants can still eventually yield rounds or cloves for replanting. At the onset of fall, if any plants have not shown signs of senescence, carefully dig the plants to examine for formation of a round. If none has formed or is immature, the plant can be returned to the ground to continue its growth cycle. In our experience in a climate with winter temperatures seldom dropping below 15°F, we simply leave these “green onion” garlic in the ground and harvest them the following summer. In colder climates, a cold frame or cool greenhouse may be required to preserve these plants. If the garlic seeds are planted early enough in the year “green onion” plants will be an exception and not the norm.
Like the seed, newly harvested garlic bulbs, including rounds, have a natural period of dormancy. Various things affect dormancy but temperature is a major factor. Warm temperatures lengthen dormancy and cold temperatures shorten dormancy. We do not have a definitive recommendation, but as a starting point we suggest waiting at least a month after harvesting the rounds before replanting them in the fall. We have not found this to be an issue for us, but if necessary temporary storage in the refrigerator could be employed to help break dormancy sooner.
Once the rounds are planted in the fall the seed-produced garlic is on a normal cycle and can be grown out just as one would with the rest of the garlic crop. Large rounds from vigorous seedlings should yield fully-developed plants and divided bulbs at harvest the following year. You can use these fully-developed plants to produce second generation seeds, though doing so severely diminishes the bulb and thus inhibits or sacrifices replication of what is essentially a new cultivar. You can also opt not to produce second generation seed from these plants and instead harvest the bulbs and cloves at maximum size for planting in the fall. This delays the next generation of seed-produced garlic until the following year, when you can use some plants for producing seed, and some for continued replication of the new cultivar via cloves. There isn’t a right or wrong way in this regard, and you may want to try some both ways. If you have a particularly promising cultivar, however, you may want to asexually replicate and preserve it before using some of its plants for a second generation of seed production.
Year 1: Remove bulbils from the umbels of selected garlic plants. Harvest the seeds in the fall.
Year 2: In January or early February begin the bleach and four week cold treatment of the seeds. In February or early March plant the seeds indoors in seed trays. As soon as feasible, migrate the seedlings to natural light and the garden bed. Depending on growing conditions, intermediate steps may include transplanting to small pots and relocating to greenhouse or cold frame. As the plants begin to senesce in late summer or early fall, harvest the rounds for replanting. Plant the rounds in the fall when you normally plant your garlic crop. If the plants have not yet formed a round by early fall, leave in the ground or, if necessary to protect from severe cold, move to a cool greenhouse or cold frame.
Year 3: Most rounds will yield fully-developed plants and divided bulbs. Harvest as you would the rest of your garlic crop, or use the plants for a second generation of seed production, or do a combination of both. In the fall, plant the cloves of particularly vigorous or promising cultivars to preserve the cultivars and increase planting stock.
Growing garlic from seed offers the opportunity to produce new cultivars that enhance and expand garlic’s genetic diversity. It also offers the prospect of more vigorous and higher yielding plants that are free of pests, viruses, and other diseases.
In subsequent generations of seed-produced plants, producing garlic seed and growing the seed into plants becomes much easier, perhaps not much more difficult than producing onion seed. Some subsequent generation seed-produced cultivars will no longer require the tedious task of bulbil removal to produce seed. The seed of subsequent generation seed-produced plants have a far higher germination rate and yield a far higher percentage of seedlings with vigor and merit. It may become routinely viable to plant seed in late winter or early spring and harvest fully-developed divided bulbs that same year. Sowing seeds directly in the ground is a viable prospect. Our early efforts at this affirm the potential.
Selective breeding by those so inclined can further refine desirable characteristics of newly developed cultivars. It may be possible, for example, to develop new cultivars that are resistant to viruses and other diseases, or a cultivar that has the large cloves of a Porcelain garlic but the complex flavors of a Purple Stripe garlic, or a Rocambole cultivar that can be grown successfully in warmer climates---and so on.
Fully developed bulbs from seed started the previous year. The largest bulb is 2½ inches in diameter. The bulbs were left to mature in the ground longer than usual since all would be replanted.
It is our hope that this article will spur others to produce garlic from seed. It is also our hope that one day Seed Saver’s Exchange members will offer true garlic seed, bulbs of new seed-grown cultivars of merit, and bulbs of garlic cultivars that are capable of producing seed without removing bulbils. It will take collective time and effort, but we think we can get there---and we think that it is very much worth getting there.
We initiated a thread on Seed Savers Exchange Forum for exchange of garlic seed reports, questions, and information exchange. Unfortunately Seed Savers Exchange subsequently eliminated all of their forums and removed all of the content. Those interested in a garlic seed information exchange forum may now wish to explore this forum:
Ted Jordan Meredith lives in Washington State and is the author of The Complete Book of Garlic: A Guide for Gardeners, Growers, and Serious Cooks. His website is at http://garlicseed.blogspot.com/
Avram Drucker lives and farms in Oregon where he grows a diversity of garlic and other alliums. His website and allium catalog are at
Etoh, T. 1985. Studies on the sterility in garlic. Allium sativum L. Memoirs of the faculty of agriculture of Kagoshima University 21: 77--132.
------. 1986. Fertility of the garlic clones collected in Soviet Central Asia. Journal of the Japanese Society for Horticultural Science 55: 312--319.
Etoh, T., and P. W. Simon. 2002. Diversity, fertility and seed production of garlic. In Allium Crop Science: Recent Advances. Ed. H. D. Rabinowitch and L. Currah.
Jenderek, M. M. 2004. Development of S1 families in garlic. In Proceedings of the XXVIth International Horticultural Congress. Advances in Vegetable Breeding.
Eds. J.D. McCreight and E. J. Ryder. Acta Horticulturae 637.
Jenderek, M. M. and R. M. Hannan. 2000. Seed producing ability of garlic (Allium sativum L.) clones from two public
------. 2003. Phenotypic characteristics of open pollinated garlic progenies. HortScience 38: 673. (Abstract).
------. 2004. Variation of reproductive characteristics and seed production in the USDA garlic germplasm collection. HortScience 39: 485--488.
Kamenetsky, R. and H. D. Rabinowitch. 2001. Floral development in bolting garlic. Sexual Plant Reproduction 13: 235--241.
Meredith, T. J. 2008. The Complete Book of Garlic: A Guide for Gardeners, Growers, and Serious Cooks. Portland, Oregon: Timber Press.
Simon, P. W. 2000. New Opportunities for an Old Crop: The Origins of Seed Production in Garlic. Mostly Garlic. (2000 Spring): 30-31, 38.
------. 2002. The Origins and distribution of garlic: How many garlics are there? The Garlic Press. (2002 Spring) 40: 1-2.
Simon, P. W., and M. M. Jenderek. 2003. Flowering, Seed Production, and the Genesis of Garlic Breeding. Plant Breeding Reviews, 23: 211-244.
Volk, G. M., A. D. Henk, C. M. Richards. 2004. Genetic diversity among