P O I N T   O F   V I E W

Triploids are fertile

This article originally appeared in the Spring 1994 issue of The Daylily Journal. Since nobody seemed to have the answers to my questions, I embarked on my own research program for hybridizing with triploid daylilies, and the ongoing results are presented here. The added-on material is italicized; also new are the links to the photographs (no photos appeared in the original article), and you can view any photo by clicking on its purple-and-underlined name. - Nick Chase

I was intrigued by Jim Brennan's comments on triploid daylilies in his article on chromosome karotypes ("The Chromosomes of Hemerocallis", second part, Daylily Journal, Spring 1992, Vol. 47 #1, p. 75), which for me raise more questions than they answer. Probably this is out of my ignorance.... somehow, I managed to miss biology courses in both high school and college, perhaps because the prospect of dismembering dead frogs did not appeal to me.... so my knowledge of biological processes is strictly that of a layperson. Jim wrote, "Because the process of meiosis [cell division for reproduction] requires the pairing of similar chromosomes and there is no mechanism allowing for the alignment of three similar chromosomes, triploid plants are not able to produce fertile reproductive cells. They are therefore sterile and unusable as parents."

I agree with the first sentence, in that triploid daylilies do not produce reproductive cells of 16-1/2 chromosomes, which can combine to form triploid children of 33 chromosomes per cell, in the way that diploid parents produce reproductive cells of 11 chromosomes or tetraploid parents produce reproductive cells of 22 chromosomes which then combine to produce diploid children with 22 chromosomes or tetraploid children with 44 chromosomes. But the second sentence, that triploid daylilies are sterile and unusable as parents, which is certainly the commonly-held belief about triploid daylilies, seems to me a nonsequitur contradicted by historical evidence. Jim goes on to write, "Rarely, such plants [triploids] can produce either haploid [11 chromosomes] or diploid [22 chromosomes] reproductive cells, but it would be highly unlikely to recover such cells or to find them producing offspring in nature."

Again, I think historical evidence tells us that "rarely" may not be as rare as commonly believed, and for this we can thank A.B. Stout. As retold by Erling Grovenstein ("That's A Good Question", Daylily Journal, Summer 1990, Vol. 45 #2, page 138) Stout made 7,135 crosses onto the known triploid H. fulva 'Europa' using diploid pollen and obtained 70 seeds (presumably all diploids). That's about a 1% success ratio of seeds to crosses, which gives a numeric value to "rarely". Certainly Stout thought it was a frequent-enough occurrence to describe the "capsule" of 'Europa' in his book (Daylilies, reprinted by Sagapress, 1986) on page 25. Erling further notes that Stout had more success with H. fulva 'Europa' as a pollen parent, obtaining 1200 seedlings when its pollen was crossed to fertile diploids. If the number of crosses was about the same - say, around 7500 - then that would give about a 15% success ratio of seeds to crosses.

Indeed, Walter Erhardt in his book Hemerocallis - Daylilies (Timber Press, 1992) on page 44 says of 'Europa', "....although it is self-sterile, since it is triploid, its pollen is fertile. This is why it became the male parent of many of the early hybrids...."

One early hybrid is MARGARET PERRY, a 1920 creation of Amos Perry's, which is a cross with H. fulva 'Europa' as the pod parent and H. fulva 'Cypriani', a diploid, as the pollen parent; this must have been a very difficult cross to make.

Another early hybrid was FULCITRINA, now no longer in commerce, a 1903 cross by Charles Sprenger and Willy Mueller using H. citrina, a diploid, as the pod parent and the triploid H. fulva 'Maculata' as the pollen parent. H. fulva 'Maculata' is also minimally pod-fertile, perhaps more so than 'Europa', and Stout illustrated it with its seed pod on page 29 (Plate 6) of Daylilies. About 'Maculata' Erhardt writes, on page 49 of his book, "....the pollen is fertile and so can be used for breeding."

So much for the supposed "sterility" of triploid daylilies when crossed with diploids.... triploid species are self-sterile. We need to be careful not to so quickly consign the knowledge hard-won by the pioneers to the dustbin of daylily history.

Most interesting in the summer of 1994 were my experiments with H. fulva 'Europa'. The pollen is definitely difficult to use, but I did get one seed with tetraploid KATE CARPENTER (Munson, 1980) as the pod parent. I had more luck setting pods on 'Europa' using tetraploid pollen; out of about 800 crosses, I got 20 seeds from 8 pods. These all were planted in late May 1995, and as of late June there were little green shoots coming up.... about a 65% germination rate overall; the seed from KATE CARPENTER as pod parent did germinate. The behavior of 'Europa' is interesting; it sets pods like crazy, but nearly all of them fall off because of mismatched ploidy. More information on this, a 1996 update, follows the horizontal rule.

About the fertility of triploids when crossed with tetraploids virtually nothing is known for sure. The only person I know of who has tried it is Stanley Saxton. I was intrigued by his catalog description of triploid GARNET ROBE (Carl Millikan, 1947) which he noted "sets seed by tets", so I wrote him for more details. He wrote back that years ago he used diploid pollen on GARNET ROBE to no avail (a result one would expect in light of Stout's experience with laying diploid pollen on 'Europa').

His experience with tetraploid pollen was more successful; the most promising seedlings came from GARNET ROBE X FAIR ANNET (Virginia Peck, 1967), the best of which was POLLY PINK, a salmon-pink which he has registered and introduced. He gave me no clue to the success rate of his crosses, or to the fertility of GARNET ROBE's pollen.

So I think I need some answers from the biologists here. Admittedly, the chromosomes in triploid daylily cells don't divide evenly in two for reproduction... but exactly how do they divide? Do they mostly split up into unusable fragments? Do they tend to divide more along "clean" lines, producing goodly numbers of cells with 11 and 22 chromosomes, with the rest in fragments? Do they divide "cleanly" but the chromosomes get scrambled in the process, rendering many of them infertile at conception time? Or does something else go on that I don't know about? Does anybody know the answers to these questions?

In answering a question on "spore-producing" daylilies in the fall Journal (Daylily Journal, Fall 1993, Vol. 48 #3, page 270), Jim Brennan says that daylily pollen grains contain three cells, two of which are sperm cells, and that the ovary contains ovules of eight cells each, one of which is the egg cell. Conception occurs, I presume, when one of the sperm cells joins an egg cell, but that doesn't tell me much about the conditions needed for successful fertilization. Must all three of the pollen cells be of the same ploidy (11 or 22 chromosomes), and all eight of the ovulian cells of the same ploidy (even if the ploidies of pollen and ovules are different) for successful fertilization to occur? Or does only the ovulian ploidy matter? Or does the ploidy of cells other than sperm and egg not matter at all? Whatever the process, experience with 'Europa' would seem to indicate that viable pollen is produced about 15% of the time, and viable ovules about 1% of the time, for crossing with diploids.

Returning to Jim's 1992 article on chromosomes, I was equally intrigued by his statement that "Occasionally, an error in the process of meiosis [cell division] results in a reproductive cell that has the diploid number of chromosomes [22 for diploid daylilies]. If this unreduced reproductive cell joins with a normal haploid cell [11 for diploid daylilies], the result is a triploid offspring...." Gee, that sounds pretty easy. Our gardens are full of tetraploid daylilies, all of which have reproductive cells with 22 chromosomes, and diploid daylilies with 11-chromosome reproductive cells. I assume there are other factors in the process, otherwise we would all be out in the garden crossing our diploids with our tetraploids and creating triploid daylilies with gay abandon. Of course, we all know what happens in real life.... the pods fall off. Perhaps Jim's next sentence holds the clue: "In the plant world, triploids are typically weak, unsuccessful organisms." That may be almost always true of daylilies, too.... perhaps in real life we can conceive triploid daylilies at will, but nature abhors triploids and aborts them shortly after fertilization.

I am sure I'm not the only person who has forgotten the ploidies of parent plants and attempted to cross tetraploids with diploids. Last summer I tried to cross HONOR BORN (Virginia Peck, 1979), a dormant tetraploid which blooms a pinkish-orange in my garden, with PASTEL DELIGHT (Currier McEwen, 1976) which I mistakenly thought at the time was a tetraploid and which grows cream with a pink blush in my garden. PASTEL DELIGHT has fertile pollen and HONOR BORN is a wonderful pod parent, and I had the poor thing loaded with so many pods, the scapes bent under the weight. Certainly fertilization had occurred and the pods grew to full size; I anticipated harvesting several hundred seeds in the fall. Only when I returned from vacation and found that all but one of the pods had shriveled and fallen off did I recheck the ploidies of the two plants and realize that PASTEL DELIGHT is a diploid. The one remaining pod, badly shriveled and hanging on by a thread, yielded three seeds. Are these seeds tetraploids, the result of accidental pollination or unreduced sperm cells from PASTEL DELIGHT, or are they triploids which nature deemed worthy enough not to abort? I can't tell, but I intend to plant them this spring and see what comes up.

There aren't many reports of successful crosses of diploids with tetraploids, but perhaps the best-known is LIGHTS OF DETROIT (Judith Weston, 1982). Weston indicates this plant resulted from using diploid BOYD HORTON as the pod parent with tetraploid FABULOUS PRIZE as the pollen parent, and that LIGHTS OF DETROIT crosses with both diploids and tetraploids, more successfully as a pod parent (Steve Webber's Daylily Encyclopedia, 1988, page 91). I see four possibilities for the ploidy of this daylily:

1. It is a "true tetraploid" caused by an unreduced egg cell from BOYD HORTON fertilized by a normal sperm cell from FABULOUS PRIZE; 2. It is a "true diploid" caused by a normal egg cell from BOYD HORTON fertilized by a superreduced (11 chromosomes instead of 22) sperm cell from FABULOUS PRIZE; 3. It is a chimera, caused by mutation of the embryo, or the fusion of two embryos, one a diploid and the other a tetraploid, into one seed which grew one plant; 4. It is a triploid which survived nature's tendency to abort triploids.

In the first case, it would seem to be a rare error that would cause an egg to have double the normal chromosomes (remember, this is not the same as colchicine-doubling of chromosomes in cell-division for growth, which is an induced process, and colchicine is a powerful chemical, to boot). An error in the second case is more believable, as cells are always dividing for reproduction and could conceivably break into 11-, 11- and 22-chromosome cells instead of the usual 22/22.... but probably still a rare occurrence. But in either case, a mutation would have to take place in the same generation.... a genetic change to cause the continual and irregular division (in case 1) or nonreduction (in case 2) of a significant number of both sperm and egg cells in LIGHTS OF DETROIT for the flower to be pollen- and pod-fertile with both diploids and tetraploids. Now that I consider to be highly unlikely.... there is no evidence of any such genetic change, or anything remotely resembling it, having occurred to date.... and, combined with the rarity of the initial nonreduction or superreduction, makes the odds zero that LIGHTS OF DETROIT is uniquely either diploid or tetraploid.

In the third case, that of the chimera, I don't know if it is possible to conceive and raise a chimera from seed.... I leave it to the biologists to definitively answer that question.... but I would think it is unlikely. Is it possible for a tetraploid embryo to mutate shortly after fertilization, and for a diploid cell to "split off" (or vice versa) and reproduce in the same seed (and plant)? Maybe, but it is no more likely to happen, I surmise, than the other possibilities for mutations I've suggested. And for two embryos of different ploidies (already a rare occurrence) to fuse into one seed and coexist when they have different chromosomal makeups I would think is totally unlikely.... nature would abort such an abomination much more readily than a triploid.

That leaves us with the triploid case. The conception of triploids by crossing diploids with tetraploids, I suspect, is a common occurrence; it is their survival of nature's desire to abort them that is uncommon..... but, I also suspect, not nearly as rare as chance mutations. So if I were a gambling man, I'd lay heavy odds that LIGHTS OF DETROIT is a triploid.

Judith Weston spoke at the March 1994 meeting of the New England Daylily Society and, because of this article, she gave an "update" on hybridizing with LIGHTS OF DETROIT. She said she had its pollen DNA-tested at the University of Minnesota. The scientists claim that this is as reliable as chromosome-counting, and that genetically, LIGHTS OF DETROIT is a tetraploid..... but Judith doesn't really believe that herself - she thinks the scientists screwed up somewhere - because the plant crosses so easily with diploids. She showed some slides of really unusual-looking tetraploid seedlings from LIGHTS OF DETROIT, mostly as a pod parent..... exotic, spoonbill- or orchidlike blooms, my kind of daylily flower! I hope she introduces some of them.

The only discrepancy I see is that the fertility behavior of LIGHTS OF DETROIT (with diploids) does not seem to match that described by Stout for triploids. On the other hand, tetraploid daylilies didn't exist in Stout's time, and he had no triploid daylily hybrids to work with.... only the species H. fulva.... so who is to say what "normal" behavior is? Because I want to experiment with triploids in my garden this summer, and because I don't want to waste time on approaches that are likely to be unproductive (do I try to put tet pollen on my enormous clump of H. fulva 'Europa', for example), I have thought about and developed a theory for what constitutes "normal" fertility behavior for triploids. I want to emphasize this is only a theory, which needs to be proved (or disproved) in the garden and in the laboratory. My theory is: Two kinds of triploids exist, according to their parentage.

The first kind includes the triploid species or near-species triploid hybrids (GARNET ROBE, for example, emerged from Carl Milliken's diploid breeding program), which for simplicity I will call S-triploids. When their cells divide for reproduction, the sperm and eggs wind up mostly with oddball numbers of chromosomes.... part of the "irregularities in sporogenesis" to which Stout refers in describing H. fulva 'Europa' on page 24 of his book. Only a few sperm and eggs have 11 or 22 chromosomes, so the S-triploids are relatively infertile.

The second kind are the modern triploid hybrids (such as LIGHTS OF DETROIT) which result from crossing diploids or triploids with tetraploids; for simplicity, I will call them M-triploids. The tetraploid parentage here is the key; I think (for reasons the biologists will best be able to explain) the already-matched chromosomes supplied by the tetraploid parent induce more regularity in their subsequent reduction (probably along the lines of the two parents), yielding considerably more sperm and eggs with 11 or 22 chromosomes and fewer with the oddball numbers. So even though M-triploid plants have 33-chromosome cells just like the S-triploids, their fertility behavior is that of chimeras, and they are relatively fertile.

The triploid children of S-triploids crossed with diploids will be S-triploids, so this would not seem to be a good area in which to hybridize. The triploid children of S-triploids crossed with tetraploids will be M-triploids, as will the triploid children of all M-triploid crosses with each other and all crosses of diploids with tetraploids. So hybridizing for triploids should probably be directed toward crossing diploids with tetraploids (if you can tolerate watching nearly all the pods fall off), M-triploids with diploids and with each other, and triploids with tetraploids. Somebody who has a "chimera" raised from seed with at least one tetraploid parent probably has an M-triploid. Similarly, someone who has raised a "diploid" from a "reduced-gamete" cross of a tetraploid parent also probably has an M-triploid.

If a large number of M-triploid hybrids were to be created, their fertility behavior would probably follow a Gaussian distribution. Some would probably cross fairly readily with tetraploids (though be considered "difficult"), but not with diploids, and would be registered by their creators as tets; others would behave inversely and be registered as dips. But the large majority should be expected to follow what I surmise is "typical" M-triploid behavior and cross to some extent both ways with both diploids and tetraploids (and with each other).... and this is the only test I can think of, short of counting chromosomes, to verify that a hand-crossed and seed-raised daylily is an M-triploid.

Saxton's POLLY PINK also crosses with both diploids and tetraploids, and by applying the same reasoning I used for LIGHTS OF DETROIT, I suspect it also is an M-triploid plant. Assuming both are indeed triploids, then Stout has demonstrated that a triploid pod parent can nurture diploid seed; Weston has demonstrated that a diploid pod parent can nurture triploid seed; and Saxton has demonstrated that a triploid pod parent can nurture triploid (and probably also tetraploid) seed. Is a triploid pod parent more likely to carry triploid seeds to maturity than a diploid or tetraploid parent because the matched ploidies provide a more "compatible" nurturing environment? Maybe the biologists can answer that question, too.... or maybe we'll just have to go out into the garden and try it.... but my initial reaction is that it probably doesn't make any difference what ploidy the pod-parent of a triploid seed is.

What's more important is that crosses be found that produce M-triploids with a higher survival rate from the tendency to abort. A century of daylily hybridizing has transformed the humble species into the round, piecrust-ruffled, picoteed beauties of today, so one can hope that sufficient interest and time will lead to the creation of beautiful triploids that produce a goodly quantity of viable triploid (as well as diploid and tetraploid) seed. I can see the triploid daylily serving as a "bridge" between diploids and tetraploids; maybe someday, we will have "seamless breeding", where diploids crossed with triploids produce triploid children which are then crossed with tetraploids, and vice versa, with perhaps the latter being more important, since nobody has yet devised a way to diploidize a tetraploid daylily.

I realize that instead of giving answers I have mostly posed a lot of questions I hope other people can answer. I also realize that the thought of coping with S-triploids and M-triploids in a daylily world which heretofore has had cleanly-separated diploid and tetraploid lines will be unsettling to many people. But I do think we should cast away the misconception that triploid daylilies are completely sterile; indeed, they may offer some interesting hybridizing possibilities. And I think we can be more sympathetic to those people who have hybridized plants they swear cross with both diploids and tetraploids; they probably aren't crazy after all.

- June 28, 1996