Polyploidy and Unreduced Gametes
Joseph C. Halinar
Published: The Daylily Journal, Vol 45, No. 4, 1990

Daylily hybridizers started to seriously breed tetraploid daylilies about 30 years ago. At that time many people thought that tetraploid daylilies would take over and completely dominate the daylily garden. Tetraploid daylilies have come a long way since then but have yet to over take the diploid daylilies. A basic understanding of polyploidy, its limitations and the difficulties of breeding at the polyploid level would have made this evident.

The tetraploid revolution occurred by the conversion of diploids to the tetraploid level by treatment with colchicine and breeding within this tetraploid pool. However, hybridizers are still converting the best diploids to incorporate new germplasm into the tetraploid germplasm pool. Colchicine treatment is a hit and miss affair and often results in unstable chimeras rather then stable true tetraploid conversions. Hybridizers in other genera have successfully used unreduced gametes to transfer germplasm between different ploidy levels and this should also be possible with daylilies. Breeding tetraploids by the aid of unreduced gametes is much easier and cleaner then treatment with colchicine. To be able to understand how to use unreduced gametes, we first must have an understanding of polyploidy and chimeras.

Daylilies have a diploid chromosome number of 22. This is referred to as the somatic chromosome number and is written as 2N = 22. The parents in a daylily cross each contribute one set of 11 chromosomes to their progeny. This is the haploid or basic chromosome number, N = 11. Each haploid set of 11 chromosomes contributed by each parent is referred to as a genome. Genomes are generally represented by a capital letter. For example, H. fulva could be FF while H. minor could be MM.

Polyploidy occurs whenever there is one or more extra genome compliments. For example, triploid H. fulva would be FFF, or 3N=33, while tetraploid H. fulva would be FFFF or 4N=44. There are two major types of polyploidy: autopolyploidy and allopolyploidy. Since triploids appear to be insignificant in daylily breeding, polyploidy will be discussed from the point of view of tetraploids.

AUTOTETRAPLOIDS. Autotetraploids have four genomes that are the same or similar enough to pair during meiosis. These chromosomes are said to be homologous, meaning the same or identical. During meiosis the four homologous chromosomes pair but fail to undergo normal disjunction and usually fail to produce viable gametes. Thus, autotetraploids usually have low fertility. Autotetraploids are produced whenever a species or a fertile hybrid is converted to a tetraploid by colchicine treatment.

ALLOTETRAPLOIDS. Allotetraploids have two sets of different genomes. If two species are crossed, they may have genomes that differ enough so that they will not pair during meiosis. For example, species A x species B will produce hybrid AB. When we look at meiotic cells in this hybrid we will see that the A and B genomes fail to pair and there will be 22 single chromosomes rather than 11 pairs. These unpaired chromosomes fail to undergo normal meiosis and the hybrid will, thus, be sterile. However, if we now double up the hybrid to a tetraploid, we will then have two sets of A chromosomes that pair with each other and two sets of B chromosomes that can pair with themselves. The chromosomes between these two genomes are said to be homeologous, that is, they are similar but not close enough to pair. Chromosome number one of the first genome is said to be homeologous to chromosome number one of the second genome, and so on for all 11 chromo- somes. Thus, when we look at this allotetraploid we see 22 pairs of chromosomes lined up during meiosis. Allotetraploids are generally fertile because they exhibit diploid like meiosis. Allotetraploids that have diploid like behavior are referred to as amphidiploids. From a practical point of view we can assume that tetraploids showing a high degree of fertility are behaving as amphidiploids. Arisumi (1) examined pollen mother cells of some colchicine induced tetraploids that were fertile and found mostly bivalent pairing.

POLYPLOID GENETICS. Each gene on a chromosome occupies a particular location called a locus. In a diploid there are two sets of chromosomes and thus each gene is present in two doses. Each gene at each locus can be present in one or more different forms called alleles. In a diploid there are only two possible alleles at each locus for any given individual even though there may be more then two alleles of that gene in the germplasm pool. For example, if 3 alleles of gene A exists within a diploid population, there are then 3 different heterozygous genotypes possible within a individual plant. These will be a1a2, a1a3 and a2a3. For four alleles the possibilities will be a1a2, a1a3, a1a4 a2a3, a2a4 and a3a4.

At the tetraploid level there are four doses of each gene at each locus. With four alleles the number of possible heterozygous genotypes at the tetraploid level is much greater then at the diploid level. Also, more then two alleles can be present, for example, a plant could have genotypes a1a1a2a3, a2a2a3a4 ... a1a2a3a4. It should be evident that at the tetraploid level there are far more possibilities for allelic interactions at a given locus.

Doubling up a diploid to a tetraploid with colchicine does not increase intra-locus allelic interactions because no new alleles are added to the genome. A diploid with genotype a1a2 becomes a1a1a2a2 at the tetraploid level.

The homeologous genomes in an amphidiploid allows for fixed heterozygosity. Fixed heterozygosity occurs when a particular locus on the first genome is homozygous dominant, for example AA, but the same locus on the second genome is homozygous recessive, aa. Because no crossing over occurs between these two genomes, all the gametes produced will be genotype Aa and all the progenies produced from crossing genotype AAaa x AAaa will be AAaa. In this case it is impossible to recover the aaaa genotype which could be a desirable genotype. Fixed heterozygosity results in phenotypic buffering because many potential phenotypes can not be expressed. Thus, segregating populations of tetraploids may exhibit far less variation than what we might otherwise expect. Both desirable and undesirable traits can become fixed.


Somatic polyploidization occurs when ever there is a disruption in the normal process of mitosis such that the duplicated chromosomes are reunited into one cell rather than dividing into two cells. If the somatic polyploidization occurs in a cell that is a precursor to other cells, then all the daughter cells will be polyploid. Somatic polyploidization can occur anywhere in the plant, but only becomes important to daylily hybridizers when it occurs in the apical stem meristem and affects the cells that will eventually form the pollen and egg cells.

Somatic polyploidization is usually brought about by treating the apical meristem with colchicine. The success rate for conversions is rather low, although some hybridizers have been fairly successful in achieving stable tet conversions. Many of the resulting plants are chimeras and eventually revert back to the diploid level. An understanding of the basic structure of the plants meristem will help in understanding why this occurs.

If we were to remove the leaves from a daylily plant we would eventually see a tight whitish cone like structure made up of immature leaves. If we were to vertically slice into this cone we would eventually see, with the aid of a microscope, a conical actively growing apical meristem, or "growing point." This apical meristem is composed of 3 cell layers which are labeled L1, L2 and L3. The L1 layer is the outermost layer that produces the epidermis cells of the plant. The L2 layer is the second layer which produces the cells between the epidermis layers, the pith cells and the germ line that will eventually produce the sex cells. The L3 layer is the innermost layer which produces most of the root structure. A good analogy is a chocolate coated vanilla ice cream cone. The chocolate coating is the L1 layer, the vanilla ice cream is the L2 layer and the cone is the L3 layer. A diploid will have a L1-L2-L3 structure of 2N-2N-2N or 2-2-2 while a tetraploid would be 4-4-4.

Depending on where the polyploidization occurs in a plant, we could get any one of three different types of chimeras. These are sectoral chimeras, periclinal chimeras and mericlinal chimeras.

Sectoral chimeras. A sectoral chimera is a chimera that occurs anywhere in a plant that does not involve the apical meristem. Sectoral chimeras are usually of little value to the daylily hybridizer as they usually fail to encompass the germ line. Sectoral chimeras are sometimes evident in petals or sepals that have mutations resulting in a different color. If we were to throw a single drop of butterscotch on our ice cream cone we would have a sectoral chimera.

Periclinal chimeras. Periclinal chimeras occur when the polyploidization occurs in the uppermost apical cell of the apical meristem such that all the resulting daughter cells in the entire cell layer are polyploid. If only the apical cell of the L1 layer became polyploid, then all further growth from this meristem would have the L1 layer polyploid and all the resulting epidermis cells and stoma cells would be polyploid while the L2 layer and the resulting egg cells and pollen would be diploid. This would be classified as a 4-2-2 chimera. Such chimeras may appear to be tetraploids, having larger stoma and guard cells but are diploid in their breeding behavior. If the apical cell of the L2 layer becomes tetraploid, then the epidermis layer will be diploid while the egg cells and pollen will be tetraploid. This would be a 2-4-2 chimera and would behave as a tetraploid when we use it for breeding.

If we were to take our chocolate coated vanilla ice cream cone and pour a large amount of butterscotch on the very top of the cone, we would then have a butterscotch coated vanilla ice cream cone, or if we were to replace the vanilla ice cream with banana ice cream, we would then have a chocolate coated banana ice cream cone. This would be equivalent to a mutation or polyploidization resulting in a periclinal chimera of the L1 and L2 layers, respectively.

Mericlinal chimeras. The active meristem is a very short crown of cells. The top most cell is considered the progenitor of all the resulting cells below it. However, cells just below this progenitor cell are also actively dividing. If the polyploidization occurs in one of the cells adjacent to the top most progenitor cell, then all the daughter cells from this cell will be polyploid but not the entire cell layer. These chimeras are generally referred to as mericlinal chimeras because they have characteristics of both periclinal and sectoral chimeras. For example, half of a stem may be polyploid while the other half is diploid. If we were to take our chocolate coated ice cream cone and pour a small amount of butterscotch near the top the butterscotch would flow down in a streak from the top down to the cone, but only part of the ice cream cone would be coated with butter- scotch.

Mericlinal chimeras are often evident in daylilies that develop variegated foliage. Some times a variegated midrib will develop in a leaf on one side of the plant and then all the resulting leaves that develop on that side of the plant above the original variegated leaf will also be variegated, but all the leaves on the other side will be green.

The most stable tetraploids occur when all three cell layers are tetraploid. When we treat a apical meristem with colchicine, we do not know where the polyploidization will take place. Most likely various cells in the apical meristem will be converted and the first flowers will produce a mixture of tetraploid and diploid gametes. New crowns produced from these treated meristems will be a complex mix of tissue types. Eventually, most plants will sort them selves out to be either complete diploids, complete tetraploids or a more or less stable periclinal chimera. Sectoral and mericlinal chimeras are generally not stable through vegetative propagation although mericlinal chimeras can sometimes be converted to a periclinal chimera if a lateral bud forms directly underneath the mericlinal chimera.

Daylilies produce new crowns each year. These new crowns form by budding off of tissue just below the apical meristem (1). Periclinal chimeras will tend to maintain themselves through this process of forming lateral buds but sometimes the L1, L2 or L3 layers will change places. Thus, a 2-4-2 plant that breeds like a tetraploid can revert to a diploid in the following year by replacement of the L2 layer with L1 tissue during the process of forming lateral buds.


Sexual polyploidization occurs when there is a disruption in the meiotic process resulting in a diploid gamete being formed rather then the normal haploid gamete. During the meiotic process each chromosome divides to form two chromatids. Thus, during chromosome pairing in diploids there are four chromatids lined up. These four chromatids separate in a two stage process and form four haploid cells. These cells then undergo further development to either form egg cells or pollen grains. Sometimes, there is a disruption in this two stage separation and subsequent cell wall formation and only two cells are formed. These cells then have the diploid number of chromosomes and when they mature they either produce egg cells or pollen grains that have the diploid number of chromosomes. These gametes are referred to as unreduced or 2N gametes.

There are basically three different types of unreduced gametes (2) with different breeding consequences. It is beyond the scope of this article to go into the detail of the different types of unreduced gametes. Suffice it to say that from a practical point of view we have to take what ever type occurs in the plant we are dealing with.

Unreduced gametes have been used successfully in many genera to transfer germplasm between different ploidy levels and there is no reason why it should not be possible to use unreduced gametes to transfer germplasm from diploid to tetraploid daylilies. The lack of naturally occurring tetraploid daylily species does not preclude the presence of unreduced gametes. The genus Lilium does not have any naturally occurring tetraploid species, but unreduced gametes are quite common and are being extensively used by polyploid lily breeders.

The production of unreduced gametes is under genetic control, usually simple recessive genes (2) but is also subject to fluctuation caused by environmental factors. Some individuals will produce a low frequency of unreduced gametes while other individuals will produced a high frequency. Most individual will either not produce unreduced gametes or do so only at a very low rate. The production of unreduced gametes on the male side can easily be checked by examining pollen grains for size. Unreduced pollen grains are larger then normal haploid pollen grains and can usually be easily distinguished from normal pollen grains with the aid of an inexpensive microscope. Detecting unreduced gametes on the female side is difficult but can be inferred if diploid x tetraploid crosses result in tetraploid progenies. High temperatures (105-110F in Lilium) at the time of meiosis will sometimes increase the production of unreduced gametes. Meiosis in daylilies occurs when the buds are about 1/2 inch long and may take 2 or 3 days to complete. Hybridizers wishing to try high temperatures to increase the frequency of unreduced gametes will need to place the diploid pollen parent into a hot greenhouse for several days sometime after the first bud is about 1/2 inch long. All the buds that were undergoing meiosis during the high temperature treatment could potentially have a increased frequency of unreduced gametes.

The possibility of using unreduced gametes in daylilies is made easier by the apparent presence of a triploid block in daylilies. When ever we make a tetraploid x diploid cross, 4N x 2N, we would expect the resulting progenies, if any, to be triploids. With triploid block all the resulting triploid embryos that might occur fail to develop. Thus, a 4N x 2N cross will or should be infertile if only haploid pollen grains are produced, but if the pollen parent produces unreduced gametes, we will then have a 2N egg cell uniting with a 2N male gametophyte with the production of tetraploid progenies.

Tetraploid hybridizers have failed to utilize unreduced gametes, possibly because they may be rare in daylilies. How- ever, a number of hybridizers have reported that they thought a particular cultivar was a chimera because it would cross with both diploids and tetraploids. More likely, these are diploid plants producing unreduced gametes. These diploid plants should be considered as potentially valuable for the transfer of diploid germplasm to the tetraploid level.

Unreduced gametes offer an easy and quick way to incorporate large amounts of diploid germplasm into tetraploid breeding lines. When doing so, the diploid parents should only be used as pollen parents. A 2N x 4N cross could produce diploid progenies from insect pollination and hybridizers who are not careful about checking the ploidy level of these seedlings could end up with a diploid breeding program masquerading as a tetraploid breeding program.

The production of unreduced gametes is under genetic control. Thus, we can cross parents that produce unreduced gametes and greatly increase the chances that the resulting progenies will also produce unreduced gametes. Crossing a diploid that produces unreduced gametes with a diploid that does not will most likely result in progenies that do not produce unreduced gametes. However, self-pollinating or sib mating these seedlings should result in about 25% of them producing unreduced gametes. Thus, hybridizers could easily transfer large quantities of diploid germplasm from diploid breeding lines to tetraploid breeding lines.

Tetraploid hybridizers who have abandoned their diploid breeding programs may wish to reconsider their breeding strategy if they make use of unreduced gametes. Tetraploid hybridizers could do substantial germplasm development at the diploid level much easier then at the tetraploid level and then use unreduced gametes to transfer that germplasm into their tetraploid breeding program.


1. Arisumi, Toru. 1972. Stabilities of Colchicine-Induced Tetraploid and Cytochimeral Daylilies. J. Heredity, vol 63(1).
2. Peloquin, S.J. 1983. Genetic engineering with meiotic mutants. Pollen: Biology and Implications for Plant Breeding, Elsevier Science Publishing Co., Inc.

Additional notes added April 1993:

This article may be difficult to understand if you don't have any background in basic botany. If this is the case and you want to understand the basics more thoroughly, then I would recommend that you read a good basic introductory college level botany text book.

Many people have a difficult time with understanding the concept of a genome. A plants genome is its total genetic content. In daylilies this consists of 11 pairs of chromosomes for a total of 22 chromosomes. These 11 pairs of chromosomes are generally represented by a capital letter, often times the first letter of the species name. Thus, H. minor could be represented as having genome MM where each M represents 11 chromosomes, one set of 11 chromosomes from the pod parent and one set of 11 chromosomes from the pollen parent.

Let us say we want to cross daylily A with daylily B. We could write this as genome AA crossed with genome BB. We could also graphically represent each chromosome in daylily A as [ so that when we look at daylily A at meiosis we would see the following:

1 [] Note: I reversed the second [ to make
the pairing more visually pleasing.
2 [] I could have showen it as [[
3 []
4 []
5 []
6 []
7 []
8 []
9 []
10 []
11 []
where each [] represents the two complementary chromosomes pairing.

For daylily B we could represent the chromosomes by + so that the paired chromosomes during meiosis would look something like the following:

1 ++
2 ++
3 ++
4 ++
5 ++
6 ++
7 ++
8 ++
9 ++
10 ++
11 ++
Each chromosome in daylily A has a complementary chromosome to pair with and the same is true for daylily B. Thus, daylilies A and B are fertile.

Now, let us cross daylily A with daylily B. The resulting genome will be AB. The question we are interested in is whether or not genomes A and B are complementary, that is, will they pair? In this example chromosome number 1 of daylily A is [ while chromosome number 1 of daylily B is +. In this case the chromosomes are different enough so that they will not pair:

[ + just can't figure out a way to pair up during meiosis.

A complicating factor occurs when the chromosomes are similar so that some pairing occurs, but not all the chromosomes are fully paired. These hybrids will show some degree of fertility depending on the amount of pairing. The more pairing the greater the fertility. For example we could make a cross where the genomes are represented by / and \. These two plants are a lot closer than the [ and + in the previous example. In this case / and \ may pair, but the pairing may not be complete for all the chromosomes.

Lets now go back to daylily hybrid AB. The genomes [ and + are just too different to be able to pair. Thus, this hybrid will be infertile. Sometimes the word sterile is used in place of infertile, but infertile is the more correct term. If we were to look at the pollen mother cells during meiosis we would see 11 ['s and 11 +'s floating around with out any pairing. These cells would not be able to successfully complete the meiotic process, so no egg or pollen grains would be produced. If, however we were to treat this hybrid with colchicine and double up the chromosomes we would then have genome AABB. The original genome is AB, but in the tetraploid we have doubled this to AB + AB = AABB. In the diploid hybrid we would see 11 [ and 11 + chromosomes. In the tetraploid we have duplicated each of the 11 [ chromosomes and also each of the 11 + chromosomes. Thus, we have 11 pairs of [[ and 11 pairs of ++ with each chromosome having a complementary chromosome to pair with. Thus, when we look at this tetraploid we see 22 pairs of paired chromosomes:

1 []
11 []
1 ++
11 ++
This plant looks like a diploid with 22 pairs, or 44 chromosomes. Thus, we call such allotetraploids amphidiploids because they behave as diploids during meiosis.

Doubling up the chromosomes in a fertile diploid often results in low fertility in the resulting tetraploids. This occurs because there are four copies for each of the 11 different chromosomes. During meiosis there are now four chromosomes trying to pair with each other and only two chromosomes can pair at any one location. The result is that these chromosomes can not separate properly and the resulting egg and pollen grains are not formed. Plants that have four copies of the same genome are called autotetraploids. For the most part autotetraploids are either infertile or have low fertility, but there are always exceptions, although not common.