Some ploidy questions

A number of forum posts have had to do with ploidy in roses, something that I would like to understand better. I understand from both roses and other flowers that higher ploidy levels are thought to lead to larger and/or “stiffer” flowers. However I do not know how consistent this relationship/correlation is. I have a few related questions along the lines of this general topic.

  1. My understanding is that in some cases different ploidy levels exist in the same species. In those cases, will roses of the same species exhibit any notable difference in flower size/form based only upon their different ploidy levels?

  2. Along the same line, when diploid roses have their chromosomes doubled artificially, does their flower size/structure change in any notable way? For example, I have an artificially doubled rose (from wichurana x banksiae). Since I only have the doubled form, I cannot compare it to the original. But if I could, would the flowers have any notable difference in size/form?

  3. I think of certain classes of roses having “general” ploidy levels associated with them. For instance, I think of polyantha roses and many ramblers as being generally diploid (although I expect that there are some exceptions) and generally also having small flowers. In these classes is the correlation between flower size and ploidy because of diploid roses generally having smaller flowers, or is the correlation merely an artifact from those classes’ ancestries (in many cases involving diploid R. multiflora and/or diploid R. wichurana)?

  4. In a contrasting example, I think of hybrid teas as generally being tetraploid, although again I understand that there are exceptions. Again, is there something about higher ploidy level (i.e. tetraploid) that is necessary for the form of hybrid teas (with their thicker petals and upright stems), or is the correlation between their ploidy and plant form simply an artifact of the ancestry of the hybrid tea class?

I once read someone suggesting that fertile triploids could be used as a bridge to bring traits from tetraploid roses to the diploid level, or vice versa. If the general ploidy levels of different rose classes are mostly an artifact of each class’ ancestry (and not inherently necessary for that class’ form), then this seems reasonable. If there is something inherent to the ploidy level that makes the characteristics of that class possible, then that suggestion might not be reasonable. For instance, could hybrid tea form actually be brought into diploid roses, or does the ploidy level of hybrid teas do something important in making hybrid teas’ form possible?

This is a topic I’ve been interested in…thanks in advance for any helpful thoughts.


Starting at the bottom, the “hybrid tea” form is borrowed from the diploid Tea roses. Have a look at ‘Isabella Sprunt’ for an example:

The correlation between flower size and ploidy holds only within a particular lineage. E.g., a tetraploid Rosa multiflora would probably have somewhat larger flowers than a diploid form of the species, but nothing like the much larger flowers of Teas and HTs that bear only one flower per stem. ‘The Bride’, a white sport of ‘Catherine Mermet’, is a diploid, but gives larger flowers than some of the triploid Hybrid Polyanthas and tetraploid Floribundas.

Thickness of petal is another character that correlates with ploidy only within a lineage (or closely related specimens). Some diploids have thick, waxy petals. Some tetraploids have thin and fragile petals.

Years ago, daylily breeders found that tetraploid selections had larger, thicker petals than the diploid originals. However, the tetraploids were too brittle, so they were backcrossed to diploids. The triploids were satisfactorily large and thick, but less brittle. After further breeding and selection, tetraploids were raised that are not brittle.

Different ploidy levels in a species can happen. I recall one case (not a rose) where the diploids and tetraploids were the same size and appearance in most of their parts. An experienced observer could distinguish them (according to what I read), but I don’t recall any mention of crossing between the two forms. If they did not cross, then one might argue that the reproductive isolation should grant them at least subspecific status.

I have read about examples in Rosa, but I haven’t seen the plants. It might happen that a tetraploid plant gives rise to a diploid seedling from an unfertilized ovum. Such a plant would resemble the tetraploid parent in general qualities, while being smaller and less fertile.

Hi Matt! I like Karl’s description. I’ve done a lot of doubling and also finding haploids in roses. The differences in thickness of plant parts, etc. due to ploidy exist, but need to be considered in reference to genetic background. For instance, diploid rugosa foliage is thicker and larger than both diploid and tetraploid polyanthas.

So, when comparing a diploid and tetraploid rose of the same genetic background, this is what we find for the tetraploid:

-larger cells at the cellular level that typically have thicker cell walls.
-less branching of the overall plant
-greater brittleness of the overall plant
-thicker tissues due to the larger cells (leaves, petals, stems)
-darker green leaves because of the thicker tissue and the light reflects off of it more versus the light radiating through the leaf
-leaflets that are wider due to wider angles from the primary and secondary veins
-larger tissues/organs
-often a bit slower growth and with less branching and therefore a tendency for less overall number of flowers.

Some old hybrid teas interestingly are triploid like ‘Tropicana’ and ‘John F. Kennedy’. I got some cuttings started and counted those. With the elevated chromosome number and awkwardness in meiosis sometimes as chromosomes pair and segregate, it is uncommon, but happens often enough to find some triploids between even crosses of two tetraploids. I suspect the nice balance in features in some genetic backgrounds between the diploid and tetraploid levels help these triploid seedlings stand out more (bit more branching than many of the tetraploid siblings, etc.).

I think that many breeders just move forward with their breeding not worrying too much about ploidy as long as they can get some seed. Triploids can be great bridges like mentioned. I think the strong tendency for triploidy in landscape roses these days is due to it being a great balance in the growth features (moderate pretty good branching with the genetic background used, still a good growth rate and flowering, etc.) as well as increased sterility so there are less fruits generally forming to help the plants basically self deadhead and keep blooming. There of course are exceptions of triploids that set a lot more hips that this may not be true for (Lemon Fizz, Carefree Spirit, etc.).

It is fun for us as independent breeders to work with bridging ploidy, working with combining diverse backgrounds, etc. to creatively find something very unique to set ourselves apart. Commercial breeders seem to have more pressure to get something that can fit a commercial niche more readily and tend to stir the same genetic pots to a greater degree. Using ploidy as a tool to help diversify the ranges of phenotypes we find I think is a great thing to try to pursue. I have some new tetraploid polys that set some seed last year and have some seedlings that are about to flower. There is much less branching than the diploid poly seedlings now, but hopefully there’ll be larger flowers and some other traits that’ll push the range of phenotypes.

Karl and David Z,
In all the the known classes of roses, lets say Rugosas, Damask, Floribundas, can we assume they are “all” the same ploidy. Can we assume that Damask are lets say “diploids”, if this is true, is there a chart to show what the classes are ?. Hope this makes sense.

David M.
The “pure” Damasks and Floribundas are typically tetraploids, though some of the older “Hybrid Polyanthas” are triploid or (rarely) diploid.

I was surprised when I learned that the so-called Russian Rugosas and the so-called Japanese Rugosas are all derived from seeds collected in a Japanese garden. I have no description of the parent(s). However, I note that among the Russian group there was at least one variety with Rugosa-like leaves that was known locally as R. cinnamomea. I take it that most of these are diploids.

BTW, Hurst (1925) wrote of the Damask group, “It is significant that a form of this tetraploid species AACC has produced several triploid forms AAC under cultivation in Holland and France, and it may be significant that the individual form examined by me is said to have been introduced from Persia, where the closely allied pentaploid species AACCE was found growing on Omar Khayyam’s grave.”

Has anyone ever artificially doubled a triploid hybrid from a diploid-tetraploid cross to create a hexaploid? That might be an interesting way of bringing a greater percentage of genetic material from a diploid into tetraploid roses more quickly, by then backcrossing the hexaploid to the original diploid parent. I doubt that hexaploid roses would have much advantage over tetraploids as garden subjects themselves, assuming they aren’t actually worse, but there’s always a chance that there might be something interesting in that to explore along the way.


1 Like

The following 2016 paper may be of interest:

Here is a paper that Dr. Kermani and her colleagues have reporting on chromosome doubling the triploid rose ‘Iceberg’. The hexaploid interestingly is soft pink.

Thanks for the reply to my question

Thanks for the link. I will admit, right off the bat, that I cringed while reading that abstract.

“In most rose species only one ploidy level is observed. However, some can have different ploidy levels. For example R. chinensis can be di-, tri- or tetraploid, while R. acicularis can be tetra-, hexa- or octoploid.”

As far as anyone can determine, Rosa chinensis Jacq. was based on a garden variety rose derived from two or more ancestral species. Even the Crimson China (Slater never saw it) carried genes derived from R. multiflora and R. luciae, among others. We really should stop confusing groups of garden plants with formal species. China roses are not Rosa chinensis, just as Rugosas are nor R. rugosa and Gallicas are not R. gallica.

And it’s not just roses that get mistreated in this way. The German Irises are not derived from Iris germanica, despite what some Catalog writers have claimed.

Is Rosa acicularis really “tetra-, hexa- or octoploid?” Did the speaker forget the diploid R. acicularis nipponensis Crépin?

Unlike so many other botanists, Hurst (1932) was able to consider more than a handful of traits at one time:
“… we have the interesting case of Rosa acicularis, which is said by different cytologists to have diploid, tetraploid, hexaploid and octoploid forms. A critical taxonomic diagnosis of the material used shows clearly, however, that the original R. acicularis of Lindley is octoploid and includes two distinct genetical species with septets AACCDDEE and BBCCDDEE, while the American hexaploid “acicularis” are either R. Bourgeauiana Crép. which is BBCCDD, or R. Sayi Schwein., which is CCDDEE. The tetraploid “acicularis nipponensis” of Willmott (non Crép.) found in the Kew collection is a subspecies of R. pendulina L., which is DDEE, while the original diploid R. acicularis nipponensis of Crépin is a subspecies of R. rugosa Thunb., which is CC.”

For those not familiar with Hurst’s shorthand, each of the letters, A through E, represents a set of 50 traits that Hurst found to hold together in species around the world. For example, The A group includes the Synstylae, Indicae and Banksiae.

Hurst’s terminology is out-of-date. Back in the 1920s, when he got started, he assumed (as almost everyone else did) that the identifying traits of a species must involve genes scattered among the seven chromosomes of the haploid set. Since then, however, we have a better (though imperfect) understanding of “supergenes”.

David, thank you for the link to that very interesting paper! The pink flower color and the apparent reduction in seed fertility are surprising results, but a lot of the other changes seem in line with those seen in the chromosome doubling of diploids. It would have been useful to know if the hexaploid’s pollen was mostly 3n (unless I missed something!), but that seems the most likely scenario. I suppose that you might get some unreduced 3n pollen from the triploid as well, but the odds would be lower.


Again please excuse my question, will all of these “ploidy” things cross with other ploids ?

David M., I think the simplest answer to that is probably “yes, they will”, although it would also be safe to add “it’s complicated”. Generally speaking, if there’s a combination you’d like to make and you can get pollen from one parent and the other is a plausibly willing recipient, then you should never let considerations of any ploidy differences completely derail any plans–those factors should only inform your approach and shape your expectations. Based on the body of evidence there are some things that you might tend to expect when crossing roses having certain disparate chromosome numbers (and those tendencies can at times be exploited to help achieve certain breeding outcomes), but there are also numerous variables at play like unreduced gametes that keep everything unpredictable and interesting.


1 Like

Thanks Stefan, most of your answer I can understand. So if I was to have a cross with, lets say “Laevigata” with “Peace” one of Kordes most outstanding achievements. It has stood the test of time. It is possible it “could” work ?

“Possible”? Perhaps. “Probable”? Probably not. Getting anything to cross with Laevigata can be “challenging”, to be gracious. There are some types which are simply mulish while others are rabbits on fertility drugs. Minutifolia, diploid, fought me for decades, until I hit on the climate, method and mates making that possible. Serafinii, a pentaploid, resists EVERY attempt, in both directions. Even when trying it with the triploid mate which made the Minutifolia offspring possible. Mixing ploidies can work and can result in some pretty remarkable creations.

1 Like

Thanks Kim, so mixing “ploidy” can/could work

You’re welcome, David. Of course it “can” work and has innumerable times.

‘Peace’ came from Meilland.
Hybrids of Rosa laevigata are known, but the one with R. banksiae [R. x fortuniana] is the best known.

I have a list of some others.

In addition, I suspect (but can’t prove) that the “Reblooming Anemone” that was growing at the San Jose Heritage garden last time I was there, is really a Rugosa x Laevigata hybrid raised by Luther Burbank. The fifth picture from the top is the typical form of cane, while the sixth looks like the Rugosa ancestry has pushed itself into expression.

Shifting dominance is not a rare phenomenon. In fact, I have another case of Rugosa traits behaving strangely.

A Rose Odyssey (1937: p123-124)
J H Nicolas

In 1933 I had found a curious sport on Margaret McGredy. The foliage strongly resembled Rugosa but the plant characteristics also leaned toward R. cinnamomea. I mentioned this fact to Sam III [McGredy] when I visited him in 1934. Sam could not account for the sport. He had never used species in his breeding. His brother-in-law, Walter I. Johnston, spoke up, “Your father did much more work with species.” We adjourned to the office, where complete hybridizing records from the early days of the firm are kept, one volume for each year, a valuable library. After several hours of research we traced the origin of Margaret McGredy to crosses of Rugosa and Cinnamomea. They were, of course, many generations back. But as these two species are in the blood stream of Margaret McGredy and all modern McGredy roses, the possibility of the sport was explained. It is an accepted fact that hybrids alone sport (pure species mutate, but rarely, if ever, sport) and can sport only within what is in them.*

Lately, the most unusual thing has happened to that sport. A sport is supposed to be a part of the hybrid compound which “took a walk”. But this sport must have carried the whole pack as it has sported again a Hybrid Tea type with a magnificent bloom much more intensely colored than the original Margaret McGredy and is a distinctly a different rose. I am planning to name it “Margaret Second”.

Also, my list of Laevigata Hybrids contains a note from Lord Penzance. In his case, the shiny leaves of Laevigata did not appear in any of the seedlings, even when Laevigata was the seed parent.

1 Like

I had a seniors moment Karl on the breeder of “Peace”, yes you are correct it was Meilland.

I had my Ginkgo biloba this morning, so I was ready. :wink:

Now, getting back to ploidy, you can think of “ploid” as “ply” as in plywood and 2-ply paper towels.

In roses, a “ploid” or “layer” is made up of seven chromosomes. These chromosomes are all different. That is an important fact to remember. So when a monoploid (one-layer) pollen tube introduces itself to a monoploid ovum, the result is a diploid cell with two layers or sets of chromosomes. Assuming that pollen and ovum are from the same or similar species, all is well. In time an embryo is formed, which grows into a fertile plant.

But if the monoploid ovum is visited by a pollen tube with two sets of chromosomes, there may be some problems … eventually.

The triploid plant that results from the cross may grow well and be very healthy: think of ‘Iceberg’. Mitosis is not a problem because the 21 (three sets) chromosomes don’t have to pair off. Just a bunch of bachelors hanging out.

When the time comes for meiosis (the reduction division that leads to new pollen and ova) the chromosomes can get tangled up looking for partners.

Instead of pairs that are easily separated when the cell divides, there may be some groups of three, each one trying to form cross-overs with the other two. As the cell divides, the three tangled chromosomes may all end up in the same daughter cell, while the other daughter cell does not get even one.

As I mentioned above, the seven chromosomes in the basic set (or ploid) are all different. And they are all needed. You may think of them as recipe cards. All are necessary for the finished product to come out right. Different versions of the recipe are allowed, as we finding when we cross a Tea rose with R. multiflora. And nothing goes terribly wrong when there are three versions of the recipe working together in the triploid ‘Iceberg’. It’s only when meiosis tries to sort 21 chromosomes into gametes with balanced recipes (sets of 7 chromosomes).

In some triploids there are enough gametes (pollen and ova) with 7 or 14 chromosomes (or nearly) to make them useful parents. Again, I point to ‘Iceberg’. But other triploids get closer to half and half, which is never a good thing for a triploid. Figuratively speaking, some of the cakes our recipe is supposed to be making will come out with too much flour and no sugar at all.