Rosa Nitida and Virginiana: Is anyone using them?

I am new to hybridizing and I have been looking to find some species roses to use. I am getting a sucker of Arkansana from Paul. Virginana and Nitida have really caught my eye. Are they worth trying? Are they hard species to work with?



I have tried them both. I love the offspring from R. virginiana. R. virginiana is tetraploid and can cross with modern roses pretty readily, especially as a male. It also for some reason tends to let warm colors (carotenoid pigments) come through to offspring from crosses with warm colored modern roses better than other species. Also, at least what I’ve been using can pass along disease resistance pretty well too. R. nitida has been more of a challenge for me and I gave up with it, but I think there is potential. Since it is diploid that has kind of hindered me. I can get crosses of it with rugosas, but it didn’t cross well with other diploids like polyanthas. I think if one would invest more at the diploid level with it or chromosome double it, there are a lot of nice traits that one can hopefully get transmitted to modern roses. R. palustris is also very very nice. What I’ve been working with is a tetraploid form (most seem to be diploid). Sincerely, David

I have 4 seedlings of (Corylus X OP) X (acicularis-tetraploid ? X OP)

Corylus is R. nitida X R. rugosa ruba. It should be a diploid. This is just a “breeding material only” type cross. Both parents, Corylus and acicularis, are single, once blooming. Both are hardy and disease resistant in northern Ohio. The cross should be a triploid but perhaps some of the pollen will be tetraploid.

I also have 1 seedling of (Corylus X OP) X mixed pollen.

My experience is that once I get the diploid species into a triploid cross, I often get some tetraplod pollen to then cross with other tetraploids.

My grafted Rosa virginiana should be arriving this month. I have not tried Rosa nitida. I sort of have to taste test species one at a time due to space limitations.

David, you should try it with tetraploids and try your luck.

Daybreaker x Rosa nitida sounds like a pretty cool start.

I’ll try R. virginiana and R. carolina this year. I also ordered several R. virginiana clones (Harvest Song and Drummer Girl) and R. carolina alba (apparently better material than R. carolina itself).

I also ordered R. x rugotida. A cross between R. rugosa and R. nitida.

Good to hear your positive experience with R. virginia, David!


According to Percy Wright, the advantage of crossing Rosa nitida with Rosa rugosa cultivars having purple red flowers is that this species eliminates the mauve tint in the progeny. For example, his ‘Aylsham’ (‘Hansa’ x Rosa nitida) is deep pink, almost red in colour with no mauve tint. This cultivar has superior flowers to the other Rosa rugosa x Rosa nitida cultivars - ‘Corylus’ and ‘Defender’. However, it’s not readily available in the U.S.

Rosa nitida, of course, can be used for breeding at the diploid level. But it may be advantageous to also bring it up to the tetraploid level. This can be accomplished by crossing it with the hexaploid Rosa acicularis. The hybrid progeny will maintain and likely improve cold hardiness and toughness of Rosa nitida. The excellent characteristics of Rosa nitida, compact form of the shrub and shiny green foliage with excellent fall colour can be selected in the progeny.

There is no advantage to use Rosa carolina instead of Rosa virginiana in a breeding program. They have the same ploidy and are similar in form and growth habit. Certainly, Rosa carolina alba is not superior to other genotypes of the species. Who says that? What’s important when selecting genotypes of species for use in breeding programs is a combination of shrub vigour and form, foliage form and colour, and flower size and colour. Flower colour alone doesn’t do it.


I did not mean to say that R. carolina alba is a superior genotype because it is white. I read somewhere that it crosses easier with other roses than plain R. carolina. There was a breeder who did not have favourable results until he switched from ‘plain’ R. carolina to R. carolina alba. Maybe this is because R. carolina alba is actually already a hybrid with something else, I don’t know.

Even though there may be no advantage to using R. carolina instead of R. virginiana, there are differences between the two species. Henry recently linked to an interesting article that showed that R. virginiana originated from a foliosa-nitidia-palustris diploid group and R. arkansana from a blanda-woodsii diploid group. R. carolina originated from a hybrid between these two groups. So, theoretically, R. carolina may have some interesting genes that R. virginiana doesn’t have. I have never seen both species. My ideas for using them are based purely on theoretical notions. We’ll see what they’ll be like in real life :slight_smile:

I am interested in your remark about crossing diploids with hexaploids to yield tetraploids. I read about this breeding route in different ‘theoretical’ sources, but I never read a story of someone who actually did it and succesfully raised a usable tetraploid. Do you know if this method actually works?


I am interested in your remark about crossing diploids with hexaploids to yield tetraploids. I read about this breeding route in different ‘theoretical’ sources, but I never read a story of someone who actually did it and successfully raised a usable tetraploid. Do you know if this method actually works?

I’d be interested in hearing about this also.

One book I read, and it was a source from the 1960s, indicated (and I’m pulling from memory here) that in normal diploid species a single pollen grain united with the ovary to form the embryo, while two pollen grains united with what became the endosperm of the seed. So the genetic ratio of 2 haploid sets for the embryo and 3 haploid sets for the endosperm. This work indicated the more the ratio deviated from this the less viable the seed.

So in the case of trying to create a triploid by crossing a diploid with a tetraploid, different ratios would result depending on which flower was acting as the seed parent. If the tetraploid was the pollen parent then the embryo would have 3 haploid set and the endosperm 5. If the diploid was the pollen parent then the embryo would have 3 haploid sets and the endosperm 4. As 3/5 is ‘closer’ to the 2/3 ratio than 3/4. In triploid crosses it would usually be likely that the diploid parent would be the better seed parent. If what I just stated has any basis in fact, the diploid/hexaploid cross would not likely be as viable as a diploid/tetraploid cross, but the hexaploid as the pollen parent would likely be more viable than the reverse.

Chris Mauchline

Randy Hudges crossed a diploid, 14, (Calocarpa) with an hexaploid, 42, (R. nutkana). This gives 7 plus 21, so yields a tetraploid. He provided me with some of the seeds. You can see how I have used them by using your Find command with Calocarpa at the link below.

Calocarpa is rugosa X chinenis.


Thanks to all for your insight. I am just getting the hang of the ploidy numbers to where it is finally making sense. I love learning all this stuff.


I’ve successfully crossed Rosa rugosa (‘Red Frau Dagmar Hastrup’) with Rosa acicularis and have one selection that I call ‘Spring Glory’. However, it grows in Victoria, B.C. so I don’t have access to it right now. ‘Carlos Dawn’ and ‘Lac La Nonne’ are other reputedly Rosa rugosa x Rosa acicularis hybrids. If I recall correctly, Robert Erskine’s ‘Caroyal’ (‘Hansa’ x ‘Lac La Nonne’) is infertile both ways and therefore perhaps a triploid, which if true would prove that ‘Lac La Nonne’ is a tetraploid. Having said that, Georges Bugnet could have been wrong with the parentage he gave for ‘La La Nonne’. He may not have been able to differentiate between Rosa acicularis and Rosa woodsii. For example, he was likely wrong in the parentage he gave for ‘Therese Bugnet’. Since ‘Therese Bugnet’ is a diploid, it would make more sense if it was Rosa woodsii (diploid) rather than Rosa acicularis (hexaploid) he stated was in this cultivar’s parentage. Furthermore, Rosa woodsii is much more common than Rosa acicularis in the area he lived (northwest of Edmonton, Alberta). And that likely would have been the native species he used in his breeding programs.

Chris, your comments are most interesting.


Hi Chris,

The endosperm/embryo ploidy and parental contributions are interesting. I don’t know the reference you are referring to or what species in particular they talk about, but this is my understanding from my training.

Typically in a pollen grain there are eventually three nuclei (all should be haploid and exact copies of each other). One is the pollen tube nucleus which typically is near the tip of the pollen tube as it goes down the style. Following are eventually two generative nuclei (they either go through the final mitosis to make two copies in the pollen tube or before depending on the species; rose should do it in the pollen tube). One of the generative nuclei fertilizes the egg to form the embryo and the other fertilizes the central cell to form the endosperm. When we consider a cross between two diploid roses let’s say- yes the embryo should be 2x and the endosperm should be 3x. The reason the endosperm is 3x is not that the there were two pollen nuclei that fertilized the central cell, but that the central cell has two nuclei of female contribution in it. They along with the one pollen generative nuclei fuse to produce 3x cells for the endosperm. So in essence in a standard diploid : diploid or even tetraploid x tetraploid cross there is a 2:1 female to male contribution of genetics in the endosperm. There are of course species with exceptions like Gaura where there is just a 1:1 female to male ratio in the endosperm because of an altered central cell, but roses should follow the general trend. Perhaps being off in the 2:1 ratio in roses due to ploidy differences from parents and/or unreduced gametes may affect if viable seed can be obtained or not, or at least rate of viable seed. If the endosperm fails then the embryo generally dies due to lack of a nutrient source.

In my potato work there is a theory called Endosperm Balance Number (EBN) (published in early 1980’s) that is very predictive of which crosses take and is based upon a 2:1 maternal to paternal ratio in the endosperm. Each species is given a value for EBN and this value ultimately in a particular cross must correspond, or be very close to, a 2:1 ratio (not just number of sets of chromosomes, but genetic factors on the chromosomes). It takes a little work going through different examples with different ploidy and EBN values before the idea seems to make a lot of sense, but basically stems from the idea of ratios of genetic contributions in the endosperm. Roses definately are more “leaky” than potatoes with different ploidies and crosses taking, at least to some degree. If people want to follow up on this idea in roses it would be great. It may be done through taking rose species with different ploidy levels in a species and cross the different ploidy levels in different directions and ways and look at endosperm development and seed set and seed viability rates. I think there may be something to rate of seed set in roses due to endosperm considerations, especially when looking at the frequencies of offspring at different ploidy levels within crosses from the work published by the French group that extracted haploids of florist roses and crossed them with different species and varieties.




Thanks for clearing up my errors, I truly appreciate it. I was bound to make errors when pulling from my memory.

I still don’t have the book in front of me, but I’ve managed to track down the author and title - It was John James and the book was ‘Create New Flowers and Plants - Indoors and Out’ published in 1964. Not a bad book and you can pick it up for as little as $2.50 (plus S&H) on the used market.

I should state that the author liked to use colchicine to double chromosomes, and also radiation to induce mutation. So I wonder if he’s still alive and if not, what killed him.

Chris Mauchline

I have worked extensively with R. virginiana as a male parent crossed with modern roses. I have had the same experience as David, in that the hybrids are quite disease resistant and can carry warm colors from their modern rose parents, unlike many hardy species. The hybrids are non-recurrent, but recurrence can probably be recovered in the second generation. I am currently working on an article describing these hybrids for the RHA newsletter. I consider this species to have great promise for breeding.

I can easily understand statements that some clones of R. virginiana and R. carolina are better than others for hybridizing. The issue is the computability of their pollen on modern roses, which ranges from excellent (often better than most pollen from other modern roses) to zero, depending on the specific clone. I have used R. virginiana ‘Harvest Song’ the most extensively, as it is slightly shorter and has attractive glandular hairs on the peduncles and fruit (rather like a light moss).

I have only worked a bit with R. nitida. Crosses with polyanthas and china/tea roses didn’t work out: few seeds, no survivors. I may try to cross it with a very fertile R. blanda clone this summer.

I have grown OP seed from Corylus as well. I never got anything remarkable. I did a bit better with seed from ‘Bella Nitida,’ which is apparently also R. nitida x R. rugosa, but has a better-looking bush and double flowers of the usual rugosa magenta. Two of the seedlings from ‘Bella Nitida’ OP were recurrent last year, in the R. rugosa fashion, so they are probably the result of ‘Bella Nitida’ X one of the many R. rugosa cultivars in the garden. My plant of ‘Bella Nitida’ is planted next to my plant of R. pendulina, a tetraploid European species. Some of the OP seedlings look like hybrids with this (‘Bella Nitida’ x R. pendulina). Such hybrids would most likely be triploid. Two of these potential pendulina hybrids made a small number of OP seeds last summer, so I am going to see what I get from them, if anything.

I am also interested in producing tetraploid breeders from R. acicularis by crossing it with diploid species. The resulting hybrid would have 3 sets of chromosomes from acicularis and 1 from the other parent (R. rugosa, R. nitida, or R. blanda in my crosses). My main problem is that, as Paul Olsen points out, R. acicularis is often confused with R. woodsii, another diploid. Last year, I finally made some R. blanda x R. acicularis crosses where I am reasonably confident that the pollen parent was actually the real R. acicularis. I have started the seeds, but is its too early for germination.

There are also tetraploid forms of R. acicularis, but getting them is the problem.

The difficulty of working with North American species is compounded by the fact that they can be very hard to tell apart and garden-collected seed is likely to be the result of hybridization with other species.

These species are also highly variable. The forms of R. carolina in commerce are selected types that make moderately tall, dense shrubs, just like the commercially available forms of R. virginiana. The forms of R. carolina native here, however, are very short and run underground so extensively that the average stems are typically about 2 feet apart. I have found the local R. carolina pollen to be very difficult to work with.

Dr. Roger Mitchell, Ferris State University, Big Rapids, Michigan (zone 5)

I was given some OP Corylus seed some years ago. I kept two seedlings. Many of the Corylus OP seedlings were prone to (some) mildew. The two I kept have never mildewed, despite spending their first summer growing under and through a mildew-prone canina seedling. They both have rugose-textured, nitida-shaped leaves. Both are singles, repeat, and set OP hips. One is white, the other is pink. The foliage of the pink looks a lot like Corylus (dark green), while the white has brighter green noticeably shiny foliage. (Based on my memory) about 60%-75% of the buds on the pink set OP hips, while not quite as many buds of the white set OP hips.