Double Fertilization
A single pollen grain lands on a stigma. Within hours, it sets in motion an event found nowhere else in the living world. Two sperm cells fuse with two different targets inside one ovule. This produces not one but two new tissues in a single stroke. This is double fertilization, the defining reproductive innovation of the flowering plants (Angiosperm Flower Structure). It explains why a wheat grain is packed with starchy flesh. It explains why a coconut sloshes with milk. It explains why every apple seed sits inside a fruit built to feed something other than itself. The phenomenon matters because it solves a problem no animal embryo faces. How does a plant guarantee its next generation gets fed before it knows how many offspring it is making? Angiosperms answer this by fertilizing a second cell alongside the egg. This instantly converts it into a nutritive tissue — the endosperm. That tissue is matched in genetic dosage to the embryo it will feed. This single mechanistic trick underlies the caloric backbone of human civilization. Rice, wheat, maize, and barley endosperm are, in tonnage, the most important food product on Earth. The big picture is one of coordinated development inside a structure smaller than a poppy seed. A pollen tube, guided by chemical cues, delivers two non-motile sperm cells to the embryo sac. One sperm fuses with the egg cell to make the zygote (2n). The other fuses with a binucleate central cell. This makes the primary endosperm nucleus (typically 3n). Two fertilization events, one pollen tube, one ovule — hence "double." This article works through the anatomy of the megagametophyte. It covers the molecular logic of pollen tube guidance and gamete fusion. It reviews the classical microscopy and genetic experiments that proved the process. It includes two worked ploidy problems and the common conceptual traps students fall into. It also traces the century-long history of its discovery. Finally, it points toward what double fertilization sets up next: Seed and Fruit Development.
- Double fertilization is unique to angiosperms: one pollen tube delivers two sperm cells, producing two fusion events in one ovule.
- Sperm 1 + egg cell → diploid (2n) zygote, the future embryo.
- Sperm 2 + central cell (usually 2 polar nuclei) → typically triploid (3n) primary endosperm nucleus, the future nutritive endosperm.
- It was discovered independently in 1898 by Sergei Navashin (in Fritillaria and Lilium) and Léon Guignard, ending decades of confusion about the "second" male nucleus.
- Endosperm ploidy and parental genome dosage matter enormously in agriculture — this is the molecular basis of the "triploid block" and the endosperm balance number (EBN) used in potato and interspecies breeding.
- Endosperm is the caloric foundation of global agriculture: wheat, rice, and maize grains are, by dry weight, mostly endosperm.
Double fertilization couples the making of an embryo to the simultaneous, dosage-matched creation of its own food supply — one pollen tube, two sperm, two fusions, two new tissues.
The core idea
Double fertilization is the process, unique to flowering plants, in which a single male gametophyte (pollen grain) delivers two non-motile sperm cells into the female gametophyte (embryo sac), and each sperm fuses with a different target cell, generating two distinct fertilization products in one continuous event. The first fusion — sperm + egg cell — is fertilization in the ordinary animal sense: it produces a diploid zygote that will divide and develop into the embryo. The second fusion — sperm + central cell — has no equivalent in animals or in gymnosperms. The central cell of a typical (Polygonum-type) embryo sac contains two haploid polar nuclei; when the second sperm nucleus fuses with both, the resulting primary endosperm nucleus is triploid (3n), and it divides to build the endosperm, a nutritive tissue that nourishes the developing embryo and, in many species, the germinating seedling. The key conceptual move is that fertilization in angiosperms is not "one sperm, one egg" but "one pollen tube, two sperm, two products." This is why seeds — unlike animal eggs — arrive pre-provisioned with a genetically matched food store rather than relying solely on maternal cytoplasmic yolk or on the mother's body throughout development. The ploidy asymmetry of the endosperm (typically 2 maternal genome copies to 1 paternal, a 2:1 ratio) is itself biologically meaningful — it sets up the parent-of-origin dosage effects central to genomic imprinting in seeds, discussed below. It is worth being precise about terms students often blur. Pollination is the transfer of pollen to the stigma; it is not fertilization. Fertilization is nuclear/cytoplasmic fusion of gametes. The megagametophyte (embryo sac) is the multicellular haploid structure inside the ovule that houses the egg and central cell; the microgametophyte is the pollen grain/tube system carrying the two sperm. Gymnosperms (conifers, cycads) undergo single fertilization only — their megagametophyte tissue that nourishes the embryo is haploid maternal tissue formed before fertilization, not a fertilization product. That contrast is precisely why double fertilization is considered a synapomorphy (shared derived trait) of the angiosperms.
Mechanism
1. Pollen germination and tube growth. A compatible pollen grain hydrates on the stigma and germinates, producing a pollen tube that grows down through the stigma and style, invading the transmitting tract. The vegetative (tube) cell controls tube growth; it carries the generative cell, which divides mitotically (either before or during tube growth) to produce two sperm cells. This trio — tube nucleus + two sperm — constitutes the mature male gametophyte in transit. Sperm cells in angiosperms are non-motile and non-flagellated; they are passively delivered, a key contrast with the motile sperm of ferns, mosses, and even the flagellated sperm retained by cycads and Ginkgo.
2. Micropylar entry and synergid destruction. The pollen tube grows chemotropically toward the ovule, guided by attractant peptides secreted from the synergid cells — two cells flanking the egg cell inside the embryo sac. In the model plant Torenia fournieri, and confirmed genetically in Arabidopsis, small cysteine-rich peptides (the LURE family, secreted by synergids) diffuse out of the micropyle (the ovule's opening) and are received by pollen-tube-expressed receptor-like kinases (e.g., Arabidopsis PRK6, and its LURE receptors). The tube enters through the micropyle (a strategy called porogamy, the norm in angiosperms), penetrates one synergid cell, and bursts, releasing its contents — the two sperm cells, the vegetative nucleus, and cytoplasm — into the degenerating synergid.
3. Gamete release and dual fusion. The two sperm cells are released essentially simultaneously and migrate the short distance to the egg cell and the central cell, which lie immediately adjacent, sharing a common cell wall region. Membrane fusion (plasmogamy) is followed by fusion of the sperm nucleus with the target nucleus (karyogamy):
- Sperm nucleus 1 + egg nucleus → zygote nucleus (2n)
- Sperm nucleus 2 + two polar nuclei of the central cell → primary endosperm nucleus (usually 3n)
Because the two sperm cells are genetically identical (both products of one generative cell mitosis) but are delivered to cells of different maternal ploidy, the two fusion products differ sharply in genome dosage despite sharing the same paternal genome.
4. Post-fertilization divergence. The zygote typically undergoes a period of quiescence, then divides asymmetrically to initiate embryogenesis (a small apical cell lineage forming the embryo proper, a larger basal cell lineage forming the suspensor). The primary endosperm nucleus, by contrast, usually divides rapidly and repeatedly without immediate cytokinesis, producing a syncytium of free nuclei in a shared cytoplasm — the nuclear type of endosperm development, seen in cereals like maize and wheat. In many eudicots (e.g., Arabidopsis), cytokinesis follows each division from the start (cellular type), and in some taxa a small chamber forms first (helobial type). Eventually, the endosperm cellularizes, proliferates, accumulates starch, protein, and lipid, and in many species is largely consumed by the growing embryo before seed maturity (as in most eudicots) or persists as the bulk of the mature seed (as in cereals and many monocots).
5. Gamete non-equivalence and imprinting. Although the two sperm cells arise from one mitotic division and appear morphologically similar, work using cell-specific fluorescent markers and single-cell transcriptomics (notably from David Twell's and Frédéric Berger's laboratories in Arabidopsis) shows the two sperm cells are not always functionally interchangeable — they can differ subtly in gene expression, and fertilization is not random with respect to which sperm fuses with which target in every species. More importantly, genes in the endosperm show strong parent-of-origin (genomic imprinting) effects: certain loci (e.g., MEDEA, a Polycomb-group gene in Arabidopsis) are expressed only from the maternal allele in endosperm, maintained by differential DNA methylation between the central cell and sperm genomes established before fertilization. This imprinting machinery is thought to have coevolved with the maternal-biased (2:1) genome ratio of endosperm, and it underlies interploidy hybridization barriers — crossing a diploid with a tetraploid frequently produces inviable seed because the maternal:paternal genome ratio in the endosperm is thrown off (the "triploid block," formalized quantitatively as the endosperm balance number, EBN, a concept developed largely from potato breeding by Stanley Johnston and colleagues in the 1980s).
Two features of the embryo sac deserve emphasis because they are so frequently mis-drawn. First, the standard "Polygonum-type" embryo sac (present in roughly 70% of angiosperm species surveyed) is seven-celled but eight-nucleate: an egg cell, two synergids, three antipodal cells, and one central cell containing two polar nuclei. Second, the antipodal cells play no role in fertilization at all and usually degenerate — a frequent source of confusion, since students sometimes assume all embryo-sac cells participate.
The evidence
The reality of double fertilization was not obvious, and it took roughly two decades of careful microscopy to settle. Early cytologists in the 1880s–1890s, using improved fixation and staining of ovule tissue, could see that a pollen tube discharged two sperm nuclei into the embryo sac, not one — but they disagreed about what happened to the second one. Wilhelm Hofmeister and later Eduard Strasburger established the general pattern of alternation of generations and gametophyte structure in seed plants, and Strasburger's students actively debated whether the "second male nucleus" degenerated, fused with a vegetative nucleus, or did something reproductively meaningful. The decisive observations came in 1898, independently, from two botanists working with different systems:
Sergei Gavrilovich Navashin (Russian Empire), working with Fritillaria (fritillary) and Lilium (lily), used serial microtome sections and iron-hematoxylin staining to follow, cell by cell, the fate of both sperm nuclei after pollen tube discharge. He documented directly, in fixed tissue at successive developmental time points, that one sperm nucleus fused with the egg nucleus while the second sperm nucleus fused with the two polar nuclei of the central cell, and that this second fusion product then began dividing to form the endosperm. This was the first unambiguous cytological demonstration that angiosperm fertilization is fundamentally double.
Léon Guignard (France), working largely with lily species in parallel, reported closely corresponding observations the same year, lending immediate independent corroboration and rapidly ending the controversy in the botanical community. What made this evidence convincing rather than merely suggestive was the quantitative and positional logic of the observations: Navashin could trace nuclear number and position through a time series — one sperm consistently paired with the single egg nucleus, the other consistently paired with the specific two polar nuclei, in section after section, across multiple individuals. This ruled out alternative hypotheses in circulation at the time (e.g., that the second sperm nucleus simply degenerated, or that it fused with only one polar nucleus, or that endosperm arose from unfertilized central-cell division alone, an idea consistent with apomixis in some species but not the general rule). The consistency of a 3n endosperm nucleus (2 maternal genome copies + 1 paternal copy) across the species he examined also matched what later chromosome counts confirmed cytogenetically once chromosome staining techniques (and eventually genetic markers) became available in the twentieth century — an early convergence between cytology and genetics. Later work extended and quantified the picture using genetic markers rather than pure cytology. In the mid-twentieth century, and more decisively with molecular markers from the 1980s onward, researchers crossed maize and Arabidopsis lines carrying distinguishable alleles (for endosperm color, isozymes, and later DNA sequence polymorphisms) and showed that endosperm tissue carries both maternal and paternal alleles in a 2:1 dosage, exactly as predicted if the second sperm truly fuses with two maternal polar nuclei rather than one. This genetic confirmation, combined with modern live-cell imaging of fluorescently tagged sperm and egg/central-cell nuclei in Arabidopsis (used to watch fusion happen in real time rather than infer it from fixed snapshots), has turned Navashin's 1898 cytological inference into a directly observed, molecularly verified process.
Worked examples
Example 1 — Tracing ploidy and allele dosage through a cross. Suppose a diploid (2n = 20) maize line homozygous for a dominant purple-endosperm allele C (genotype CC) is crossed as the female parent by a diploid line homozygous recessive c (colorless), used as the pollen donor.
- Female gametophyte: egg cell is haploid C; central cell's two polar nuclei are each haploid C, so the central cell is "CC" in dosage.
- Male gametophyte: both sperm cells carry allele c (they are mitotic siblings of the same generative cell).
Apply the two fusions:
| Fusion | Genotype | Ploidy |
|---|---|---|
| sperm (c) + egg (C) | Cc | 2n zygote → embryo |
| sperm (c) + polar nuclei (C, C) | CCc | 3n endosperm |
The embryo is Cc — heterozygous, diploid, and (if C is fully dominant) it would appear purple if embryo tissue expressed the pigment, though in maize the visible kernel color comes from the endosperm/aleurone, not the embryo. The endosperm is CCc, a 2:1 maternal:paternal allele dosage. This is precisely the xenia effect long used by classical maize geneticists (including in Barbara McClintock's era) as visible, immediate proof of double fertilization in a single generation: the paternal allele's phenotypic contribution is visible in the endosperm of the very seed produced by that cross, not delayed to the next generation as with ordinary diploid embryo traits. Now push the ploidy further, mirroring real potato-breeding problems. Cross a tetraploid female (4n = 48; each polar nucleus therefore diploid, 2n) with a diploid male (2n = 24; each sperm haploid, n).
- Zygote: n (sperm) + 2n (egg) = 3n (triploid embryo) — often weak or inviable.
- Endosperm: n (sperm) + 4n (two diploid polar nuclei, 2n + 2n) = 5n.
The "normal," viable endosperm dosage ratio in most species is maternal : paternal = 4 : 1 (i.e., 2 polar nuclei genome-equivalents × 2 = the standard central cell contributes twice the egg's genome number, and by convention the balanced ratio across the whole endosperm genome is maternal:paternal = 2:1 when both parents are the same ploidy). When parental ploidies are mismatched, that ratio is thrown off — here it becomes far more maternal-heavy than 2:1 — and endosperm development fails or is abnormal, aborting the seed even though a triploid embryo formed. This is the cytological and genetic basis of the triploid block, and it is why potato breeders quantify species compatibility using the Endosperm Balance Number (EBN) rather than raw chromosome number: EBN estimates each parent's "effective" genomic contribution to a functional 2:1 endosperm, and crosses are predicted to succeed only when maternal EBN : paternal EBN = 2:1, regardless of actual ploidy.
Example 2 — Reasoning through an ecological/agronomic scenario. A wheat breeder notices that a particular cross between two hexaploid bread wheat lines (Triticum aestivum, 2n = 6x = 42) yields plump, well-filled grain, while a cross of the same maternal line with a tetraploid durum wheat pollen donor (Triticum durum, 2n = 4x = 28) yields shriveled, underdeveloped grain even though pollen tubes clearly reach the ovules and fertilization is confirmed cytologically. Reasoning: in the compatible hexaploid × hexaploid cross, both polar nuclei and the fertilizing sperm carry the same ploidy background (each gamete is 3x = 21), so the endosperm's maternal:paternal genome ratio is the standard 2:1 (polar nuclei 3x+3x=6x, sperm 3x, ratio 6:3 = 2:1) — balanced, and endosperm development proceeds normally, filling the grain. In the hexaploid × tetraploid cross, the sperm genome is 2x, while the polar nuclei remain 3x each (6x total), giving an endosperm ratio of 6:2 = 3:1, more maternal-biased than the balanced 2:1 the endosperm developmental program expects. The imprinting and dosage-sensing machinery that normally scales endosperm proliferation to a 2:1 input is disrupted, endosperm cellularization and starch-filling falter, and the grain shrivels — the same triploid-block-style logic as in the potato case, now explaining a real bread-versus-durum wheat breeding obstacle.
Common misconceptions & traps
| Misconception | Why it's wrong | Correct framing |
|---|---|---|
| "Double fertilization means two sperm fertilize the egg." | Confuses the two separate targets. | One sperm fertilizes the egg; the other fertilizes the central cell. Two different fusions, two different products. |
| "The endosperm is just extra embryo tissue." | Endosperm is a distinct tissue with its own genome dosage and developmental program, genetically triploid (typically) versus the diploid embryo. | Endosperm and embryo are sibling products of the same fertilization event but are genetically and developmentally distinct tissues. |
| "All angiosperm endosperm is triploid." | Ploidy of the central cell varies among lineages; some species have more or fewer polar nuclei fused into it. | 3n is the common/typical outcome (Polygonum-type, 2 polar nuclei), but other embryo-sac types (e.g., Oenothera-type with 1 polar nucleus, or Fritillaria-type variants) give different endosperm ploidies. |
| "Gymnosperms also do double fertilization." | Gymnosperm "female gametophyte" tissue nourishing the embryo is haploid maternal tissue formed before fertilization, not a fertilization product. | Double fertilization (with a genuine second fusion event forming the nutritive tissue) is a synapomorphy of angiosperms, though a partial/vestigial "double fertilization" without endosperm formation has been reported in a few gymnosperms like Ephedra, a genuinely contested edge case worth flagging. |
| "The two sperm cells are randomly interchangeable and functionally identical in all respects." | They are mitotic siblings and often behave equivalently, but molecular and some cytological evidence shows subtle expression differences and non-random fertilization patterns in some species. | Treat them as genetically identical but not necessarily developmentally equivalent in every lineage — an active research area, not settled dogma. |
The scientist behind it
Origins · 2 minSergei Gavrilovich Navashin and how this was discovered
Where this matters today
Agriculture and grain quality. The endosperm is the harvested product of the world's major cereal crops. Its dosage-sensitive, imprinting-controlled growth program is a direct target of plant breeding: understanding EBN and the triploid block lets breeders predict which interploidy or interspecies crosses will produce viable, well-filled seed, guiding wide-hybridization strategies in potato, wheat, and other staples.
Hybrid seed production and seedless fruit. Manipulating fertilization outcomes underlies commercial seedless fruit and hybrid seed technologies; triploid seedless watermelon exploits exactly the fertilization-dosage mismatch discussed above (a diploid × tetraploid cross yields largely inviable seed embryos inside a fruit that still develops).
Genomic imprinting research. Plant endosperm is a major experimental system for studying genomic imprinting generally, informing comparative work on imprinting in mammalian placenta — a case of convergent evolution of parent-of-origin gene regulation in two independently evolved nutritive tissues (endosperm and placenta) that both mediate maternal-offspring nutrient transfer.
Pollen tube guidance and crop fertility engineering. The discovery of LURE peptides and their receptors has opened a route to engineering or diagnosing fertilization failure in crops, relevant to improving seed set under heat and drought stress, a growing concern for food security under climate change.
Conservation and reproductive biology of wild species. Understanding double fertilization and its ploidy sensitivities informs conservation breeding programs for polyploid wild plant populations, and helps explain natural reproductive isolation between plant species of different ploidy, a factor in speciation and invasive-species risk assessment.
Ask the tutor: How does genomic imprinting in the endosperm actually silence or activate specific maternal versus paternal alleles at the molecular level?
Is double fertilization found in any plants besides angiosperms?
No plant group performs double fertilization in the full angiosperm sense — two fusions producing both an embryo and a genetically distinct nutritive endosperm tissue. Gymnosperms nourish their embryos with haploid maternal megagametophyte tissue that formed before fertilization, not a fertilization product. A partial exception is the gymnosperm genus Ephedra, where researchers (notably William (Bill) Friedman) have documented a second fertilization event that produces a diploid tissue, but this tissue does not become a functional nutritive endosperm as in angiosperms — it is generally interpreted as an evolutionarily related but non-homologous or partially homologous phenomenon, and its evolutionary significance remains actively debated.
Does the embryo or the endosperm develop first?
Typically the primary endosperm nucleus begins dividing before the zygote does. In many species there is a lag of hours to a few days before the zygote undergoes its first division, while the endosperm nucleus proceeds rapidly through free-nuclear (syncytial) divisions. This head start makes developmental sense: the endosperm needs to establish a nutrient-supplying tissue early, since the embryo will depend on it (directly or indirectly, depending on species) for nutrition throughout early embryogenesis and often through seed maturation.
Why do some mature seeds (like beans) have almost no endosperm, while others (like wheat or coconut) are almost entirely endosperm?
This reflects two different seed strategies. In many eudicots (e.g., beans, peas), the endosperm is transient — it is produced, but the developing embryo consumes nearly all of it during seed maturation, storing nutrients instead in swollen cotyledons; the mature seed is therefore mostly embryo. In cereals (wheat, rice, maize) and some other monocots, the endosperm persists and becomes the dominant seed tissue at maturity, providing the starchy reserve consumed during germination (and by humans, as flour and grain). Both patterns start from the same double-fertilization event; they diverge only in how much endosperm survives to seed maturity.
What is xenia, and how does it prove double fertilization genetically?
Xenia is the visible effect of the pollen (paternal) genotype on endosperm phenotype within the very seed produced by a cross, rather than waiting for the next generation as with ordinary diploid embryo traits. Classic examples include maize kernel color and starch texture, both controlled by endosperm genotype. Because the endosperm's genome includes a paternal contribution (from the second sperm fusion) alongside two maternal genome copies, a dominant paternal allele can visibly alter kernel color or texture immediately. This immediate, dosage-sensitive paternal effect is only explicable if a second, independent sperm–central cell fusion delivers a paternal genome directly into the endosperm — strong genetic corroboration of Navashin's cytological findings.
How does a pollen tube "know" where to go inside the pistil?
Pollen tube guidance proceeds in stages. Initial growth through the style is guided by general chemical and structural cues in the transmitting tract. Near the ovule, guidance becomes highly specific: the two synergid cells secrete small diffusible peptides (LURE peptides, first characterized in Torenia fournieri and later in Arabidopsis) that are detected by receptor-like kinases on the growing pollen tube tip (such as Arabidopsis PRK6). This synergid-secreted attractant gradient draws the tube precisely to the micropyle, ensuring it enters the correct ovule and delivers its two sperm cells to the right target cells rather than wasting them on a degenerated or already-fertilized ovule.
What happens to the antipodal cells and the synergid that isn't destroyed?
The three antipodal cells, positioned at the opposite (chalazal) end of the embryo sac from the micropyle, generally play no direct role in fertilization; in most species they degenerate before or shortly after fertilization, though in a few taxa they proliferate transiently. Of the two synergid cells, only one is typically penetrated and destroyed by the discharging pollen tube; the second, "persistent" synergid usually degenerates soon afterward as well, though its precise timing and any residual signaling role are still studied, since synergids continue producing attractant signals that must be shut off to prevent additional pollen tubes ("polyspermy" at the ovule level) from being attracted to an already-fertilized ovule.
Can double fertilization occur without producing a viable seed?
Yes — fertilization (the fusion events themselves) and successful seed development are separable. As the ploidy-mismatch (triploid block/EBN) examples show, both fusions can occur exactly as expected, yet the resulting endosperm's genome dosage may be so unbalanced relative to the embryo's needs that endosperm development fails, starving the embryo and aborting the seed even though double fertilization was cytologically completed. This distinction — successful fertilization versus successful seed set — is central to plant breeding practice and to understanding natural reproductive barriers between species or cytotypes.