Several weeks ago, there was some discussion on the Rose Hybridizers Forum about the ploidy of roses used in breeding including some of the contemporary Rosa bracteata hybrids from Ralph Moore. David Zlesak mentioned that he had done chromosome counts on both 'Star Dust' and 'Tangerine Jewel' using the traditional root tip squash method. David determined that 'Star Dust' is a tetraploid, while 'Tangerine Jewel' is a triploid. Both are descendants of Ralph Moore's 1999 Bracteata hybrid 'Out of Yesteryear', which served as the pollen parent. This information opened the door to more questions about the nature of chromosome stabilization in complex hybrids and specifically, how R. bracteata behaves through multiple generations of breeding.

At this point, David generously offered to do root tip squashes on more of the Bracteata hybrids so that we might get a clearer picture of the cytology of these roses. Fortunately, I had in my personal collection all of the named Bracteata hybrids Ralph Moore had released to date and could obtain samples from them for testing. Also included was Moore's important breeding rambler "0-47-19" [1], which would prove useful when we later did chromosome counts on some of my own 2004 seedlings. I submitted the samples to David in three lots over a period of a few weeks and David did the tests as his schedule allowed. Here are the results of our first group of tests, including 'Tangerine Jewel' and 'Star Dust', which David had already done counts for:

We decided that since there was such a curious range of results that we should also include some of the parents of the Moore Bracteatas, including "1-72-1" ('Little Darling' X 'Yellow Jewel'), 'Sequoia Gold' ("Lemon D" X 'Gold Badge'), and 'Joycie' ("1-72-1" X 'Gold Badge') in the hopes that knowing their ploidy might help show us where the varied chromosome counts were coming from. I included 'June Laver' in the samples, because we were testing a few seedlings from my cross of 'June Laver' X 'Out of Yesteryear'. Also sampled were seed parents 'Sheri Anne' and 'Twilight Skies', and seedlings from my crosses of 'Sheri Anne' X 'Out of Yesteryear', 'Twilight Skies' X 'Muriel', and "0-47-19" X 'Out of Yesteryear'. Although 'Out of Yesteryear' was now a confirmed triploid, we believed that it was producing a high percentage of 1x and 2x pollen, and so the inclusion of some of my own seedlings would hopefully provide more information about its behavior in hybridizing. Here is what we learned from this test group:

One other cultivar was sampled; a seedling from 'June Laver' X 'Muriel' named "93-03-03", which proved unwilling. There are four seedlings from this cross that all behave like tetraploids, with slow, stocky growth and "thick" plant parts. None of these have flowered yet, but they have been kept for observation and possible inclusion in further breeding, assuming they have fertility.

'Muriel'

So, what do these chromosome counts tell us? First of all, we know R. bracteata is a diploid, with 14 chromosomes. I can find no evidence to suggest that specimens with higher chromosome counts were ever encountered. The pollen parent, 'Guinée', is a tetraploid climbing Hybrid Tea that as far as I can see, behaves as a tetraploid in breeding, giving two sets of chromosomes to its progeny. Therefore, it was presumed at the beginning of this study that 'Muriel' would turn out to be a triploid, which is in fact not the case. 'Muriel' is a tetraploid with four full sets of chromosomes, alternatively written as 4x=28. How does 'Muriel' behave as a breeder? Our tests on three of Muriel's immediate descendants show that it is not acting as a stable tetraploid all the time. Both 'Pink Powderpuff' and 'Out of Yesteryear' are confirmed triploids (3x=21). This is quite a mystery when you consider that both parents of 'Out of Yesteryear' are tetraploids. (We do not know the ploidy of 'Lulu', the Hybrid Tea seed parent of 'Pink Powderpuff', and although there is a temptation to assume that it is a tetraploid like most of its kin, we dare not, knowing what we now know about fertility and ploidy.) 'Star Magic' is the only named sibling to 'Out of Yesteryear' and it is a tetraploid. We know little of its fertility except to say that Kim Rupert reports having used it somewhat in breeding.

Note that the seedlings coded "33-03-03" and "33-03-05" ('Twilight Skies' X 'Muriel') were both determined to be a tetraploids. Two others from that cross turned out to be triploids. We know that the seed parent 'Twilight Skies' is a triploid with three sets of chromosomes, and yet it is highly fertile as a seed parent. "33-03-03" was selected from the cross because it had good coloring, a large flower (about 15 petals), and extremely healthy, thick foliage. It was also discovered that it sets seed readily when a mid-season bloom was allowed to pollinate and mature, and so it will be incorporated into my breeding program in 2005.

In the process of our study, I asked David for his opinion about what was happening when triploids arose from crosses of two tetraploids, and this was his response: "I was surprised last spring to discover one of my advanced selections is triploid, although it had tetraploid parents. So it seems possible to get spontaneous triploids from mating two tetraploids, although probably rare. Unexpected ploidy levels may be coming from tetraploids where uniform and consistent chromosome pairing during meiosis is not regularly occuring. For instance, among the chromosome-doubled polyanthas I generated there was pretty low fertility in general. They are autotetraploids instead of allotetraploids. All four sets of chromosomes theoretically should recognize each other well and be capable of pairing. Sometimes the four compliments may pair two and two, sometimes three form a trivalent and there's one loner, sometimes all four have parts of their chromosome arms pairing and join to form a tetravalent." "Joining of chromosomes in pairs of two generally allows for the most uniform separation of chromosomes to the daughter cells at the end of Meiosis I and is referred to as disomic (di=two, soma=body) pairing or inheritance (no trivalents, tetravalents, univalents...). The parent in essence behaves like a diploid in meiosis. During meiosis disomic pairing promotes balanced gametes and therefore often greater fertility than polyploid parents that do not have stabilized disomic inheritance and have variable chromosome numbers in gametes. Selection for fertility over generations often goes hand in hand with selecting for stable disomic inheritance. For instance, modern wheat is a hexaploid, but has very stable disomic inheritance. There are three "ancestral genomes" that are thought to have made up modern wheat and they each contribute two sets of chromosomes which preferentially pair with each other. The ancestral genomes of wheat are not necessarily very different. There is a gene in wheat that was found on chromosome 5 in the B ancestral genome that controls disomic pairing. When the gene isn't there or is not functional, homologous chromosomes across the ancestral genomes can pair. Imagine the chaotic pairing that can happen during meiosis within a hexaploid without disomic inheritance and the frequency of unbalanced gametes and reduction in fertility that could occur." 'Out of Yesteryear'

Specifically regarding the creation of 'Out of Yesteryear', David offered this statement:

If this is true, then it leaves the door open to questions about the dynamic nature of chromosome and gene behavior in roses. We must also now ponder the likelihood that there have been, and likely still are, many more fertile triploids in breeding lines than we previously suspected. Let's take 'Out of Yesteryear' as an example. We have confirmed that it is a triploid and yet it is highly fertile as a breeder. In fact, my personal experiences in working with this cultivar have shown that it is often far more fertile than many other roses I have worked with, known tetraploids included. It should be noted, however, that it is much easier to use successfully as a pollen parent than as a seed parent. This is likely due to the fact that a rose produces far more pollen cells than egg cells, and so there is simply a greater number of fertile 7 and 14 chromosome gametes created in the pollen.

Plants are "motivated" to reproduce sexually for the sake of their survival, and so there is a tendency for plants of mismatched, unbalanced chromosome sets to periodically produce fertile gametes with complete chromosome sets. Since all the genes roses have are spread over a set of 7 chromosomes, the reduction mechanism, when working efficiently, favors the clean division of whole sets of 7. As Ralph Moore has said to me on more than one occasion, "The rose will find a way".

The popular perception to date has suggested that with only a few exceptions, triploids are quite infertile. Three exceptions that come to mind are 'New Dawn', 'Iceberg', and 'Buff Beauty', all three of which are moderately fertile. At this point, I think it would be wise to recognize the possibility that chromosome counts likely play a smaller role in determining fertility in roses than we thought. While I understand the desire of hybridizers to think carefully about matching parents of stable/similar ploidy, I think that it would be a hindrance to discard a potentially wonderful mating strictly because of ploidy issues. As we can see, highly fertile plants can result from crosses of mixed ploidy.

One interesting case in point is an example of modern Rugosa breeding at Sequoia Nursery. Years ago, Ralph Moore bred and released the extraordinary Rugosa hybrid 'Linda Campbell', which came out of a cross of 'Anytime' X 'Rugosa Magnifica'. While we do not know for certain the ploidy of either of the parents, we acknowledge the fact that first generation Rugosa hybrids are often troubled by complete sterility. The assumption has been that unbalanced chromosome counts in gametes have been the main barrier to fertility in these instances. However, there is a sibling to 'Linda Campbell' that is currently employed in the Moore breeding program, named "Magseed", which is extremely fertile as a pollen parent and somewhat fertile as a seed parent. My own tests with "Magseed"have shown that most any fertile seed parent will readily accept "Magseed" pollen and produce plenty of germinable seed.

Let's go back an examine some of the data from our ploidy tests. The samples focused mainly on 'Out of Yesteryear' and its progeny, since there have been a good number of Moore hybrids using it, and seedlings from my own breeding program as well. Studying this cultivar allowed us to concentrate on the behavior of a highly fertile triploid in breeding. Two important breeding plants played a role in the creation of four of the Moore Bracteata hybrids: 'Joycie' and "1-72-1", the latter being a sibling to 'Rise 'N' Shine'. Both 'Joycie' and "1-72-1"are confirmed tetraploids. The two registered offspring of "1-72-1"X 'Out of Yesteryear' are 'Lemon Pearls', and 'Star Dust'. Both of these are tetraploids. This piece of information does not, however, allow us to conclude that 'Out of Yesteryear' is producing 2x pollen exclusively, since a larger sample of the cross would be needed to determine this. The cross of 'Joycie' X 'Out of Yesteryear' produced two named cultivars: 'Amber Gem' and 'Tangerine Jewel', both which tested as triploids. This suggests that 'Out of Yesteryear' is not producing just 2x pollen, since it seems likely that 'Joycie' is giving 2x eggs. Therefore, we must examine other crosses to see what 'Out of Yesteryear' is doing.

This is where testing a selection of my own seedlings has been useful. First, there is the cross of 'June Laver' X 'Out of Yesteryear'. Three seedlings were tested and one is a tetraploid, two are triploids. While we cannot be absolutely certain that 'June Laver' is consistently producing egg cells with 2 full sets of chromosomes, it is probably reasonable to make that assumption. If that is the case, then we can see that 'Out of Yesteryear' is probably producing pollen with both 1 and 2 sets of chromosomes.

'Precious Dream'

Another of the recently released Hybrid Bracteatas from Ralph Moore is his beautiful 'Precious Dream', bred from a cross of 'Orangeade' and 'Out of Yesteryear'. While this cross likely produced both tetraploids and triploid offspring, 'Precious Dream' is indeed a tetraploid. It is fully fertile as both a seed and pollen parent and I have employed it extensively in the 2004 breeding work. Perhaps next year we can test some of its offspring to determine whether or not it is acting as a fully stabilized tetraploid. I have seen a small sampling of seedlings using this cultivar in crosses with David Austin hybrids and I am convinced that it has great promise as a breeder of mannerly, compact shrubs with OGR style blooms.

Lets look at one of the other crosses I submitted for testing, code number "42-03", using Ralph Moore's "0-47-19"* as the seed parent and 'Out of Yesteryear' as the pollen parent. This cross provided me with many seedlings in 2004, from which fifteen plants were selected. Again, none of these have yet flowered, but they were all selected and potted on for further evaluation and possible inclusion in the breeding program.We have already established the fact that "0-47-19"is a diploid, and is probably producing gametes with one set of chromosomes. If 'Out of Yesteryear' is producing both 1x and 2x pollen, then we could expect both diploids and triploids to appear in the progeny of such a cross. Samples from seven seedlings were submitted. Of the seven, four were indeed triploids, one was a diploid, and most surprisingly, two were tetraploids! Up to this point we had concluded that 'Out of Yesteryear' was producing both 1x and 2x pollen, and yet somehow two tetraploids were obtained from a diploid x triploid cross. I think this is the single most remarkable piece of information to come out of our study. Either "0-47-19" is producing some egg cells with two sets of chromosomes, or 'Out of Yesteryear' is producing pollen that is a mix of 1x, 2x and 3x. The question now, are the tetraploids coming from a 2x egg and 2x pollen or a 1x egg and 3x pollen? This knowledge paints an even stranger picture of rose cytology than we had previously imagined. Can we no longer be certain of the outcome of any particular cross, no matter how predictable it may seem to be?

One of my other hybrids made in 2002 was the 'June Laver' X 'Out of Yesteryear' cross. We have confirmed 'June Laver' as a tetraploid, which does not surprise us, as the work of George Mander and others has suggested that it behaves as a tetraploid. In fact, 'June Laver' is a fairly poor grower on its own roots, taking several years to build up to a substantial plant. In roses, tetraploids often do not have the same vigor as their triploid and diploid relatives. David comments on this issue:

Three seedlings of the 'June Laver' X 'Out of Yesteryear' cross were sampled for our study. One turned out to be a tetraploid and the other two were triploids. Again, this is evidence that suggests that if 'June Laver' is consistently producing 2x egg cells, then the mixed ploidy of the offspring must have arisen from the application of 'Out of Yerteryear's mixed 1x and 2x pollen.

When I was informed of these results, it occurred to me that the tetraploid of the three was fairly identifiable by its growth traits when compared to its siblings. The two triploids were faster growing plants, less double flowers with lighter petal substance. The tetraploid, much like 'June Laver' was slower to build up growth, had thicker petals and many more of them, better color, and had a "sturdier" look about it. As David has suggested, perhaps it is possible to select for specific ploidies when culling seedlings. Can we make assumptions that the diploids are often going to have superior vigor over the tetraploids in the same cross? Will the seedlings of higher ploidy often appear sturdier and have better-looking blooms? It's probably best not to weigh in too heavily in this matter, but it would help explain how selection over the decades has favored the selection of tetraploid modern hybrids, assuming bloom quality and overall "sturdiness" is the major criteria for selection.

"174-02-03"

Other seedlings tested included "174-02-02", and "174-02-06" both from a cross of 'Sheri Anne' X 'Out of Yesteryear' and both of which proved to be tetraploids. (This result for "174-02-02" was not a surprise to me, since it was a slower grower, with an overall "huskiness" to it and very full blooms with thick petals.) We also counted Ralph Moore's as-yet-unnamed "Bracteata Moss Climber", a vigorous climbing Moss rose that blooms with soft yellow "Austin-like" blooms. This beautiful rose came from a cross of "Lemon D" X 'Out of Yesteryear'. It has a spectacular Spring bloom followed by sporadic rebloom and then another immense display in the Fall. Like its pollen parent, it too is a triploid and is highly fertile as both a seed and pollen parent, giving seeds that germinate easily. At this time we do not have access to samples of "Lemon D" to test, but at some point in the future we may be able to include it in our study.

Finally we have one seedling of mine that was selected for its extreme vigor and ability to set seed, from a cross of "Everbloom One" X 'Out of Yesteryear'. This is a glossy-leaved climber with clusters of single, 2.5 inch white blooms, sometimes showing a tint of peach. The seed parent, charmingly titled "Everbloom One", is a Ralph Moore rose that is a sibling to 'Renae', a rampant climber classed as a Climbing Floribunda. Both are cluster-bloomers and nearly thornless. David confirmed that 'Renae' is diploid; perhaps "Everbloom One" is as well? Knowing that the "Everbloom One" X 'Out of Yesteryear' seedling is a diploid, I am more inclined to include it in breeding this coming season. Perhaps it would be worthwhile to pursue a "diploid only" breeding line to see what could be done.

At this point I think the underlying message here is that as hybridizers, we can be very flexible in our thinking with regard to the cytology of the roses we are working with. We can say with a fair bit of certainty that over the decades, as breeders selected sturdier (although not necessarily more vigorous) plants with better petal substance, etc., they were tending to select for polyploidy. In roses, this has resulted in a largely tetraploid population, particularly in the majority of highly bred roses like the Hybrid Teas. When the hybridizer introduces species (or other highly dissimilar roses) into a breeding line, then the outcome is less predictable than one might think. In the case of 'Muriel', we have a tetraploid that is producing gametes with varied chromosome counts (almost certainly a mix of at least 1x and 2x), resulting in offspring whose chromosome count is not predictable, regardless of the chromosome count of the other parent.

One generation further we have 'Out of Yesteryear', a triploid bred from two tetraploids. It is my understanding that most hybridizers try to avoid creating triploids because it is commonly believed that triploids suffer from severely limited fertility or total sterility. There are known exceptions which we discussed earlier, such as 'New Dawn'. Now we have 'Out of Yesteryear' which is a proven breeder with extremely high fertility, especially as a male parent. Looking at the next generation of offspring from 'Out of Yesteryear', we examined the un-named Moore hybrid, the "Bracteata Moss Climber". This too is a triploid, and yet it is proving to be a highly fertile plant in breeding. The mauve Miniature 'Twilight Skies', which I used in breeding in 2003, has also turned out to be a triploid with high fertility. What does this mean to the experimental hybridizer? If the Bracteata group is any indication, ploidy plays a smaller role in fertility issues than we thought it did. If roses like 'Out of Yesteryear' can be as fertile as they are in spite of their chromosome counts, then it would appear that selecting for fertility in seedlings from mixed ploidy crosses is more worthwhile than avoiding making such experimental crosses.

The breeding of new rose cultivars is a fascinating process from the perspective of both the artist and the scientist. I believe that like many creative endeavors, both of these views must share control of the work, and the blending of the two is what allows for originality and accomplishment. Our understanding of rose cytology should enhance our work, not place boundaries on the creative thinking that has made the work of men like Ralph Moore so spectacular, so innovative. When contemplating a cross, allow the the question in your mind to be "Where might this lead me?" rather than "What obstacles will I encounter?"

1. "0-47-19" was bred in 1947 and came out of a cross of R. wichuraiana and 'Floradora'. Not commercially available.

2. The breeder listed in the Moore records as "Lemon D" is a cross of 'Little Darling' X 'Lemon Delight', and has never been commercially released.