{ "id": "seed-anatomy-and-storage", "url": "https://coffearium.org/entries/seed-anatomy-and-storage/", "license": "CC-BY-4.0", "title": "Seed anatomy and storage", "summary": "The coffee seed as a living seed – embryo, endosperm and silverskin inside the parchment – and why it survives partial drying but not cold storage when dry, so coffee diversity is kept in field genebanks.", "domain": "botany", "kind": "Concept", "stage": "plant", "topics": [ "science" ], "topicSections": {}, "lang": "en", "status": "verified", "reviewer": "Coffearium verification pass", "lastVerified": "2026-10-22T00:00:00.000Z", "updated": "2026-10-22T00:00:00.000Z", "added": "2026-10-22T00:00:00.000Z", "brief": [ "The coffee seed is a tiny embryo, 3–4 mm long, embedded in a hard endosperm that stores the food for germination.1", "Coffee seeds have intermediate storage behaviour – they tolerate partial drying but are injured by cold when dry – so they fit neither orthodox nor recalcitrant seeds.3,1", "Arabica seed viability falls quickly after 4–6 months at ambient temperatures.1", "In a two-year storage study, washed seeds kept in their parchment stayed viable for more than a year; all other stored coffees lost viability within 6 months.6", "Because the seeds cannot be stored for long, coffee genetic resources are kept mainly as living trees in field genebanks; cryopreservation is the long-term alternative.1,10" ], "quickFacts": [ { "label": "Seed shape", "value": "Plano-convex, furrow on the flat side" }, { "label": "Arabica seed size", "value": "10–18 × 6.5–9.5 mm" }, { "label": "Embryo", "value": "3–4 mm, axis and two cotyledons" }, { "label": "Storage tissue", "value": "Endosperm; cell walls mainly mannans" }, { "label": "Storage behaviour", "value": "Intermediate" }, { "label": "Long-term conservation", "value": "Field genebanks; cryopreservation" } ], "aka": [ "Coffee seed anatomy", "Coffee seed storage" ], "linkTerms": [ "coffee seed", "seed storage behaviour", "intermediate storage behaviour", "seed viability", "cryopreservation", "field genebank" ], "related": [ "coffee-cherry-anatomy", "germination-and-seedlings", "fruit-development", "peaberries-and-bean-shapes", "coffea-arabica", "coffea-canephora", "liberica-and-excelsa", "ethiopian-landraces", "drying", "porosity-and-cell-structure" ], "contested": [ "Lowest water content tolerated by *C. liberica* seeds: about 0.24 g of water per g of dry weight in the work of Ellis and colleagues, about 0.16 g/g (whole seeds dried at 15 °C) in a 1999 Brazilian study; both are compiled in the same 2006 review.1" ], "corrections": [ { "date": "2026-10-22T00:00:00.000Z", "change": "Storage study of 2008 described correctly (washed, natural and semi-dry coffee stored; only washed coffee kept in parchment, all other coffees lost viability within 6 months).", "reason": "The text said only hulled washed seeds were compared with parchment coffee; the abstract names three processing types." }, { "date": "2026-10-22T00:00:00.000Z", "change": "Survival below 0 °C narrowed to survival at −20 °C at about 0.10–0.12 g/g (Ellis et al. 1990); the 5 °C study dated (1979).", "reason": "The review gives survival at other subzero temperatures at other water contents; the value is specific to −20 °C." }, { "date": "2026-10-22T00:00:00.000Z", "change": "Conference paper without venue or year (former source 11) removed; the field-genebank problems now rest on the 2006 review.", "reason": "Owner ruling – conference papers without venue and year are not valid sources." }, { "date": "2026-10-22T00:00:00.000Z", "change": "Summary no longer says the seed does not survive freezing.", "reason": "Seeds dried to the right water content survive −20 °C and liquid nitrogen (sources 1, 8)." } ], "sources": [ { "id": "s1", "title": "Coffee seed physiology", "authors": "Eira, M. T. S.; Amaral da Silva, E. A.; De Castro, R. D.; Dussert, S.; Walters, C.; Bewley, J. D.; Hilhorst, H. W. M.", "year": 2006, "publisher": "Brazilian Journal of Plant Physiology 18(1), 149–163", "url": "https://www.scielo.br/j/bjpp/a/zZgRS3Mbwg8KjW8R9bn9X3L", "doi": "10.1590/S1677-04202006000100011", "tier": "A", "quote": "Coffee seeds are now considered to have storage behavior defined as intermediate", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s2", "title": "Cytology, biochemistry and molecular changes during coffee fruit development", "authors": "De Castro, R. D.; Marraccini, P.", "year": 2006, "publisher": "Brazilian Journal of Plant Physiology 18(1), 175–199", "url": "https://www.scielo.br/j/bjpp/a/hsZwcNvkPXHF97CkYgKvfKq/abstract/?lang=en", "doi": "10.1590/S1677-04202006000100013", "tier": "A", "quote": "Thereafter, the perisperm gradually disappears and is progressively replaced by the endosperm (true seed).", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s3", "title": "An Intermediate Category of Seed Storage Behaviour? I. COFFEE", "authors": "Ellis, R. H.; Hong, T. D.; Roberts, E. H.", "year": 1990, "publisher": "Journal of Experimental Botany 41(9), 1167–1174", "url": "https://academic.oup.com/jxb/article-abstract/41/9/1167/594496", "doi": "10.1093/jxb/41.9.1167", "tier": "A", "quote": "a third category of storage behaviour intermediate between those of orthodox and recalcitrant seeds", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s4", "title": "Quantitative estimation of seed desiccation sensitivity using a quantal response model: application to nine species of the genus Coffea L.", "authors": "Dussert, S.; Chabrillange, N.; Engelmann, F.; Hamon, S.", "year": 1999, "publisher": "Seed Science Research 9(2), 135–144", "url": "https://www.documentation.ird.fr/hor/fdi:010018287", "doi": "10.1017/S096025859900015X", "tier": "A", "quote": "Seed desiccation sensitivity was studied in nine species of the genus Coffea by measuring seed viability", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s5", "title": "Relationship between seed desiccation sensitivity, seed water content at maturity and climatic characteristics of native environments of nine Coffea L. species", "authors": "Dussert, S.; Chabrillange, N.; Engelmann, F.; Anthony, F.; Louarn, J.; Hamon, S.", "year": 2000, "publisher": "Seed Science Research 10(3), 293–300", "url": "https://www.documentation.ird.fr/hor/fdi:010023556", "tier": "A", "quote": "Consequently, mature seeds are shed at the beginning of the following dry season.", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s6", "title": "The Storage of Green Coffee (Coffea arabica): Decrease of Viability and Changes of Potential Aroma Precursors", "authors": "Selmar, D.; Bytof, G.; Knopp, S.-E.", "year": 2008, "publisher": "Annals of Botany 101(1), 31–38", "url": "https://academic.oup.com/aob/article-abstract/101/1/31/92513", "doi": "10.1093/aob/mcm277", "tier": "A", "quote": "Wet-processed coffee was stored either as parchment coffee, where the endocarp remained around the beans or as hulled beans.", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s7", "title": "Physiological quality of seeds of arabic coffee cultivars stored for a period of two years", "authors": "Alixandre, R. D.; Lima, P. A. M.; Alixandre, F. T.; Krohling, C. A.; Fornazier, M. J.; Guarçoni, R. C.; Almeida, T. F. R.; Dias, R. S.; Anastácio, A. S.; Alexandre, R. S.; Ferreira, A.; Lopes, J. C.", "year": 2024, "publisher": "Brazilian Journal of Biology 84, e285879", "url": "https://biblioteca.incaper.es.gov.br/digital/bitstream/item/4790/1/physiologicalquality.pdf", "doi": "10.1590/1519-6984.285879", "tier": "A", "quote": "Storing arabica coffee seeds for 24 months results in a drop in the germination percentage.", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s8", "title": "New determinants for tolerance of coffee (Coffea arabica L.) seeds to liquid nitrogen exposure", "authors": "Dussert, S.; Engelmann, F.", "year": 2006, "publisher": "CryoLetters 27(3), 169–178", "url": "https://www.documentation.ird.fr/hor/fdi:010050496", "tier": "A", "quote": "obtained with coffee seeds, i.e. 89%, a value which was not significantly different from the initial viability percentage.", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s9", "title": "Variability in response of seeds to liquid nitrogen exposure in wild coffee (Coffea arabica L.)", "authors": "Vasquez, N.; Salazar, K.; Anthony, F.; Chabrillange, N.; Engelmann, F.; Dussert, S.", "year": 2005, "publisher": "Seed Science and Technology 33(2), 293–301", "url": "https://www.documentation.ird.fr/hor/PAR00000418", "tier": "A", "quote": "seed viability after cryopreservation varied from 0 to 74% among the accessions studied", "abstractOnly": true, "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s10", "title": "Global Conservation Strategy for Coffee Genetic Resources: Summary for ITPGRFA Stakeholders", "authors": "Bramel, P.; Krishnan, S.; Horna, D.; Lainoff, B.; Montagnon, C.", "year": 2017, "publisher": "Global Crop Diversity Trust, Bonn", "url": "https://www.croptrust.org/fileadmin/uploads/croptrust/Documents/GCCS_Summaries/Coffee__3_.pdf", "tier": "B", "tierNote": "International fund for crop diversity; summary of a survey of coffee genebanks. Tier B (owner ruling, 22 Oct 2026); its numbers would need a second independent source, so it is used here only for statements without numbers.", "quote": "Among the genebanks surveyed most accessions are conserved in field collections", "accessed": "2026-10-22T00:00:00.000Z" }, { "id": "s11", "title": "The Grind – Particles and Particularities", "authors": "von Blittersdorff, M.; Klatt, C.", "year": 2017, "publisher": "In: Folmer, B. (ed.), The Craft and Science of Coffee, Academic Press, ch. 13, pp. 311–328", "url": "https://eld.bl.uk/catalog/018161895", "doi": "10.1016/B978-0-12-803520-7.00013-X", "tier": "B", "quote": "cutting of the chaff or silverskin from inside the fold of the coffee beans", "accessed": "2026-10-22T00:00:00.000Z" } ], "markdown": "## What a coffee seed is made of\n\nThe green coffee bean is a seed. It is plano-convex: one side flat with a lengthwise furrow, the other side rounded.1 In *C. arabica* it is 10–18 mm long and 6.5–9.5 mm wide.1 Inside the fruit, each seed sits in a hard shell, the endocarp, which becomes the parchment after drying.1 The fruit layers around it are described in [Coffee cherry anatomy](/entries/coffee-cherry-anatomy/).\n\n
\n
\n
    \n
  1. Parchment (endocarp): the hard shell around the seed. It strongly slows germination.1
    2. \n
  2. Silverskin (spermoderm): the thin seed coat, a remnant of the perisperm.1 Part of it stays inside the fold.11
    3. \n
  3. Furrow: the lengthwise groove on the flat side.1
    4. \n
  4. Endosperm, hard outer region: living storage tissue with thick cell walls.1
    5. \n
  5. Endosperm, soft inner region.1
    6. \n
  6. Embryo axis: hypocotyl and radicle, close to the convex surface.1
    7. \n
  7. Cotyledons: two small, heart-shaped seed leaves.1
    8. \n
  8. Endosperm cap: the endosperm in front of the radicle tip.1
    9. \n
\n
A coffee seed in its parchment. A: cut across. B: cut lengthwise through the embryo. Point at a part, its number or its name to highlight it. Schematic, not to scale.
\n
\n\nMost of the seed is endosperm, the tissue that is later roasted.1 It is alive, with a hard outer region and a soft inner region.1 The embryo is small, only 3–4 mm long.1 It has an axis and two cotyledons, and it lies close to the convex surface.1 The embryo carries few reserves of its own. It depends on the endosperm for nutrients.1\n\nThe silverskin is the thin seed coat, also called the spermoderm.1 According to a coffee science textbook chapter, part of it sits inside the fold of the bean and is set free in grinding.11\n\n## A seed that is still alive\n\nA seed is viable when it can still germinate. Coffee seeds stay alive through harvest and processing.1,6 How they sprout is covered in [Germination and seedlings](/entries/germination-and-seedlings/).\n\nViability does not last long. Arabica seeds lose it quickly after 4–6 months at ambient temperatures.1 The parchment makes a difference. In a 2008 study, washed, natural and semi-dry processed coffees were stored for two years; the washed coffee was kept either in parchment or hulled.6 Seeds kept in parchment stayed viable for more than a year; all the other coffees lost viability within 6 months.6 The authors suggest that this loss of viability contributes to the fading aroma of old green coffee.6\n\n## Why coffee seeds are hard to store\n\nSeed scientists sort seeds by how they store. Orthodox seeds are shed dry and tolerate further drying.1 Recalcitrant seeds are shed wet and die when dried.1\n\nCoffee fits neither group. A 1990 study suggested a third, intermediate category for it.3 Coffee seeds survive partial drying but not complete drying.1 When dry, they are also injured by cold.3 In a 1979 study, moist arabica seeds stored at 5 °C lost their viability at once.1 Cold storage of dry seed, the usual seed-bank method, therefore does not keep coffee for long.1\n\n## How coffee diversity is kept\n\nBecause the seeds are hard to store, coffee germplasm is kept in field collections: living trees in field genebanks.1,10 Field collections need land and labour, and they are exposed to environmental hazards and pathogens.1 According to the 2017 global conservation strategy for coffee, collections share their material mainly as seed.10\n\nThe long-term alternative is cryopreservation: storing seeds in liquid nitrogen.1 For seeds that cannot be stored in the usual way, it is the only method available for long-term conservation.1 According to the same strategy, some coffee genebanks already keep part of their collection this way.10 The Ethiopian collections are described in [Ethiopian landraces](/entries/ethiopian-landraces/).\n\n\n\n## Endosperm and embryo in detail\n\n**Origin of the tissues.** Up to mid-development, the young fruit consists mainly of pericarp and perisperm.2 The perisperm then gradually disappears and is replaced by the endosperm, the true seed.2 The endosperm is liquid at first and hardens as the fruit matures.2 The silverskin is described as a remnant of the perisperm.1 How the fruit grows is covered in [Fruit development](/entries/fruit-development/).\n\n**Reserves.** The main reserves of the seed are storage proteins, sucrose and complex polysaccharides.2 The endosperm cell walls consist of cellulose and hemicelluloses, mainly insoluble mannans.1 The lateral endosperm is extremely hard because the mannan is laid down as very thick cell walls.1 The endosperm cap, in front of the radicle tip, has thinner walls.1 These cell walls are what roasting later breaks open; see [Porosity and cell structure](/entries/porosity-and-cell-structure/).\n\n**Embryo.** The embryo has an axis and two adherent, heart-shaped cotyledons.1 Polyembryony and empty seeds have been reported at a frequency of 1.2%, in a 1941 study cited by the 2006 review.1 The endocarp drastically inhibits germination, so it is usually removed before germination tests.1\n\n## Desiccation and cold\n\nThe 2006 review gives seed water content on a dry-weight basis: grams of water per gram of dry matter (g/g).1 The figure compiles the values from the 2006 review.\n\n
\n
\n
Water contents at which coffee seeds were found to survive, lose half their viability or die, in g of water per g of dry weight, as compiled in a 2006 review.1
\n
\n\n**How dry.** Ellis and colleagues gave about 0.09 g/g as the lowest water content tolerated by *C. arabica* seeds, 0.10–0.12 g/g for *C. canephora* and 0.24 g/g for *C. liberica*.1 Their 1990 study tested four arabica cultivars after hermetic storage for up to 12 months.3 Most seeds of one cultivar withstood drying to 5–6% moisture content (wet basis), while the other three were much more sensitive.3\n\n**Species differ.** A 1999 study modelled desiccation sensitivity in nine *Coffea* species.4 The water content at which half of the initial viability is lost (WC50) ranged from 0.05 g/g in *C. pseudozanguebariae* to 0.38 g/g in *C. humilis*.1,4 A follow-up study linked this to climate.5 In simulations from rainfall records, seed development lasted about as long as the wet season, so mature seeds are shed at the start of the next dry season.5 The mean number of dry months that seeds must survive after shedding was significantly correlated with their desiccation sensitivity; seed water content at maturity was not.5\n\n**Cold.** Arabica seeds stored at 5 °C with 0.15–0.50 g/g lost their viability at once (1979 study); the review calls this an apparent predisposal to chilling injury.1 Ellis and colleagues found survival at −20 °C only at about 0.10–0.12 g/g.1 In the 1990 study, seeds lost viability faster when cool or below 0 °C and dry than when warmer or moister – the reverse of orthodox seeds.3\n\n## Reference values\n\n| Parameter | Value | Source |\n|---|---|---|\n| Arabica seed size | 10–18 mm long, 6.5–9.5 mm wide | Eira et al. 20061 |\n| Embryo length | 3–4 mm | Eira et al. 20061 |\n| Lowest water content tolerated | *C. arabica* ≈0.09 g/g; *C. canephora* 0.10–0.12 g/g; *C. liberica* 0.24 g/g | Ellis et al., in Eira et al. 20061 |\n| WC50, nine *Coffea* species | 0.05–0.38 g/g | Dussert et al. 19991,4 |\n| Arabica viability at ambient temperature | Falls quickly after 4–6 months | Eira et al. 20061 |\n| Short-term storage, arabica | Hermetic, 10 °C, about 10–11% moisture (fresh-weight basis), ≈0.11–0.12 g/g (our calculation) | Ellis et al., in Eira et al. 20061 |\n| Short-term storage, *C. canephora* | 13–15% water content (basis not stated), 10–15 °C | Eira et al. 20061 |\n| Storage at −20 °C | Most seeds of six arabica cultivars survived 30 days, hermetic, at about 0.12 g/g | Eira et al. 20061 |\n| Drying before cryopreservation | 0.20 g/g, ≈16.7% wet basis (our calculation) | IRD protocol, in Eira et al. 20061 |\n| Survival after liquid nitrogen, best protocol | 89% | Dussert & Engelmann 20068 |\n\n## Viability in storage\n\nA 2024 Brazilian study stored seeds of ten arabica cultivars for 24 months.7 The seeds were pulped, fermented for 24 hours and dried to 35 ± 1% moisture (wet basis).7 They were sealed in 0.10 mm plastic and kept at 25 ± 2 °C.7 The parchment was removed for the germination test.7\n\n
\n
\n
Germination of arabica seeds of ten cultivars at harvest and after 24 months of storage in sealed plastic at about 25 °C. Where both values are close, they are given together: at harvest · after 24 months.7
\n
\n\nAt harvest, germination was 87–99%.7 After 24 months it ranged from 2% (Japi) and 9% (Caturra IAC-476) to 99% (Catuaí IAC-62).7 The authors conclude that 24 months of storage lowers germination.7 Catuaí IAC-62 is the one cultivar whose germination rose, from 87% to 99%.7\n\nThe 2008 storage study adds the effect of the parchment.6 It also found that glucose and fructose fell slightly and glutamine fell clearly during storage, and that these changes did not follow the loss of viability.6 Re-imbibed seeds turned bluish green, which the authors attribute possibly to oxidised chlorogenic acids reacting with amino compounds.6\n\n## Cryopreservation\n\n**The protocol.** The IRD protocol (Institut de Recherche pour le Développement, Montpellier) dries seeds to 0.20 g/g, then freezes them in two stages.1 A slow cooling step at 1 °C per minute comes first, then immersion in liquid nitrogen; the seeds are rewarmed fast and rehydrated slowly.1\n\n**What decides survival.** A 2006 study found that survival depends on the quality of the seed lot and on the rewarming regime.8 Slow rewarming in ambient air was highly damaging.8 Drying at 81% relative humidity gave much higher survival than the 78% used until then.8 With precooling in a dry-ice bath for 25 min and rewarming in a 37 °C water bath for 30 min, survival reached 89%, not significantly different from the initial viability.8\n\n**Genotypes differ.** A 2005 study froze seeds of 30 wild arabica accessions from the CATIE field genebank in Costa Rica.9 Seed viability after cryopreservation varied from 0 to 74% between accessions.9 The embryos survived well; the losses were due to damage to the endosperm only.9\n\n## Open questions\n\n- **Why storage in parchment helps.** The 2008 study shows the difference; the abstract opened for this entry does not give the mechanism.6\n- **Cultivar differences in longevity.** In the 2024 study, germination after 24 months ranged from 2% to 99% between cultivars; no study opened for this entry explains the spread.7\n- **Cryopreservation at genebank scale.** Survival differs widely between accessions and depends on the endosperm.9 According to the 2017 conservation strategy, field collections still hold most accessions.10\n\n## Key studies\n\n
\n

Ellis, Hong & Roberts, 1990 – An intermediate category. Stored seeds of four arabica cultivars at different moisture contents and temperatures for up to a year. Dry, cool seeds lost viability faster than moist, warmer ones, so coffee behaved neither as an orthodox nor as a recalcitrant seed. The authors proposed a third, intermediate category.

\n

Journal of Experimental Botany 41(9), 1167–1174 · doi:10.1093/jxb/41.9.1167

\n
\n\n
\n

Eira et al., 2006 – Coffee seed physiology. A review of the coffee seed: structure, development, germination, desiccation tolerance, storage and cryopreservation. It compiles the water-content limits of several species and the storage protocols in use.

\n

Brazilian Journal of Plant Physiology 18(1), 149–163 · doi:10.1590/S1677-04202006000100011

\n
\n\n
\n

Selmar, Bytof & Knopp, 2008 – Storage of green coffee. Stored washed, natural and semi-dry processed arabica for two years, the washed coffee in parchment or hulled, and followed viability with a tetrazolium test, sugars and amino acids. Seeds in parchment stayed viable for more than a year; all other coffees lost viability within 6 months.

\n

Annals of Botany 101(1), 31–38 · doi:10.1093/aob/mcm277

\n
\n\n
\n

Dussert & Engelmann, 2006 – Surviving liquid nitrogen. Showed that seed-lot quality and the rewarming regime decide whether arabica seeds survive cryopreservation. Drying at 81% relative humidity, dry-ice precooling and fast rewarming gave 89% survival.

\n

CryoLetters 27(3), 169–178

\n
\n\n
\n

Vasquez et al., 2005 – Wild arabica in liquid nitrogen. Froze seeds of 30 accessions from the CATIE field genebank. Survival ranged from 0 to 74% between accessions, and the damage was in the endosperm, not the embryo.

\n

Seed Science and Technology 33(2), 293–301

\n
" }