American Rose Annual 45:114-118 (1960)
The Inheritance of Color in Hardy Roses
Percy H. Wright

Saskatoon, Saskatchewan, Canada

SINCE 1932 when I first placed pollen on Rosa macounii, one of the three native rose species of Saskatchewan, on pistils of the hardiest of the Hybrid Rugosa, Hansa, I have been concerned with the inheritance of color in roses. Because the roses which I was breeding were largely diploids, it follows that the inheritance of color in the Hybrid Teas and their relatives, which are in general tetraploids, will be different, and thus my observations may not necessarily have universal application. However, clearly any basic study of how color is inherited in roses will have to begin with diploids.


CybeRose note: Although blue pigments have been isolated from some roses, the "blueing" of roses is largely the result of co-pigmentation.

In spite of the fact that there is no really blue rose, a blue pigment is present in many roses. Indeed, this pigment although it his never been isolated from red pigments is a bit of a nuisance, for it fades less rapidly than either the red or the yellow pigments. This in turn means that many pink and red roses fade disagreeably, turning an unattractive violet or magenta color as they age.

Whether or not a rose fades agreeably seems to depend upon the amount of the blue pigment present in the petals when they are still folded in the bud. However, genes which control the rate of fading of the red and yellow pigments also enter into the total picture. If the red or yellow is nonfading, or relatively slow-fading, obviously a rose flower may retain its original good color throughout its life.

Hansa is the most common of cultivated roses in Saskatchewan, a variety with remarkable hardiness, hardy even to 50 or 60 degrees below zero, and yet everblooming. However, it is a flower of poor color and poor form.

Hansa is noted for a strong tone of violet when it first opens, and for a rapid fading of the red pigment which soon intensifies the violet coloration. Its original blend of color is probably merely a strong intensification of the ordinary "rose" pigment found in most rose species. Proportionately, in relation to the red, there seems to be little or no extra blue color.

From the cross of Hansa and R. macounii sprang Mary L. Evans, which has a pleasant pink-rose color as it opens, but which fades just as quickly as Hansa.


CybeRose note: Extensive research has shown that 'Harison's Yellow' and 'Persian Yellow' are triploids, not tetraploids.

Soon after my original cross, I put pollen of Harison's Yellow on pistils of R. macounii and got a number of seedlings. These were pollen-fertile in spite of the fact that one parent is tetraploid and the other diploid. One of the seedlings was a salmon-pink bicolor that combined the yellow of the one parent with the pink of the other, although both colors were relatively pale. The remainder were pure white.

That pure white appeared in these seedlings was unexpected, for surely one would have expected all these seedlings to be a salmon-pink. However, apparently, Harison's Yellow is heterozygous for its yellow color, and R. macounii heterozygous for its pink color. These white seedlings, as it chanced, did not inherit yellow from the one parent, or pink from the other and so had to be white.

It is more usual for roses which are heterozygous of a color to pass down faint tinges of color than no color at all. In fact, we always tend to get a superabundance of blush-pink varieties in breeding new roses.

Later, I placed pollen from Harison's Yellow on pistils of Rosa spinosissima altaica, the pure white rose from the Altai mountains of central Asia, and from this cross I got many new varieties. They ranged from pure white through various tints of pale yellow to the definite yellow of which two were eventually named Hazeldean and Yellow Altai. In these, the yellow is just a mite less deep than in Harison's Yellow. These two roses probably represent the maximum for yellow that can be expected to descend from the stock of yellow genes in Harison's Yellow to any progency. The seedlings from this cross produced so many more whites than yellows or creams that it suggested to me that these genes were present in only one of the two sets of chromosomes present in all tetraploids, and that even in the set which did have them, they were heterozygous, that is have come down from only "one side of the house." Speculation on the antecedents of Harison's Yellow is always interesting. Apparently this rose had more white ancestors than yellow ones.


Inevitably, if we want to breed new yellow roses, we'll find the task easier if we use at least one parent that is homozygous for yellow. In the case of tetraploids, this standard in practice will have to be relaxed to the extent of demanding homozygosity in only one of the two chromosome sets.

Among hardy roses, where shall we get parent varieties of such a constitution? Harison's Yellow, as already demonstrated, is certainly not homozygous for yellow. In all probability Persian Yellow (Rosa foetida persiana) is not homozygous either.

It is hard to tell about Persian Yellow, which has so rare fertile pollen grains that it has had few direct descendants. I once got a single seedling from putting pollen of Persian Yellow on Rosa macounii which developed into a tiny-flowered, weak-growing plant with blossoms having only the slightest tinge of yellow. At the Central Experimental Farm, Ottawa, the Hybrid Rugosa, Agnes, was originated by pollinating R. Rugosa with Persian Yellow. Agnes is an apricot-tinted bicolor, but the yellow element in it is certainly not deep enough to indicate that Persian Yellow is homozygous for yellow. Apparently, too, the original seedling which began the Pernetiana race as a result of using the pollen of Persian Yellow on a Hybrid Perpetual did not have enough yellow to justify calling it a bicolor. Only by taking advantage of segregation was it possible to obtain a true bicolor in Soleil d'Or, and true yellows in later descendants. However, considering the numerous yellow-flowered Hybrid Teas which trace their ancestry to Persian Yellow, it would today be surprising indeed if there were not several that were homozygous for yellow. These can hardly be regarded as useful for breeders of hardy northern shrub roses, though, because Hybrid Teas or parents introduce altogether too much tenderness into their descendants to make the first generation generally hardy enough for use in severe climates.

To repeat, if any rose exists that is homozygous for yellow, it most likely to be found in the Pernetiana group. Perhaps one or other of the famous breeders of Hybrid Teas in Europe, if not in North America, could tell us whether the numerous "progeny tests" which they have made have revealed the existence of a variety homozygous for yellow. If there is none, in all likelihood one could be created by undertaking a special project.


If, in the field of the hardy roses, I were to undertake such a project I would intercross the various descendants of Harison's Yellow with R. spinosissima altaica, or backcross them to Harison's Yellow. For instance, my Yellow Altai, crossed with its parent Harison's Yellow, could easily produce a rose as deeply yellow as Harison's Yellow itself, if not as deeply yellow as Persian Yellow. Further segregation just might result in a variety, not only with deep yellow flowers, but also homozygous for yellow.

That is, if we interbreed heterozygous roses many times, we stand a chance of getting homozygous roses. The laws of inheritance among diploids show that if we cross two individual plants that are heterozygous for a certain gene, the progeny, in the long run, will include 25 per cent that are homozygous for that gene. Another 50 per cent will be returns to the ancestral constitution of heterozygosity for that gene, and the remaining 25 per cent will completely lack that gene. Thus they will be homozygous too, but for absence of the gene.

I once crossed Mary L. Evans, which is Hansa by R. macounii and so almost surely a diploid, with Hazeldean, my best descendant of the R. spinosissima altaica by Harison's Yellow cross, and thus a tetraploid. There resulted so many dwarfs and plants with deformed flowers that I realized that I had run into still another form of in-compatibility. Triploids in apples seem to be "happy", but my experience in roses is that they are certainly anything but "happy" as a general rule. All the roses from this cross were single, despite the semi-double flowers of both parents. In color, all were wild-rose pink.

CybeRose note: The alternative view is that Hazeldean is heterozygous for the dominant "non-red" trait found in R. foetida (except bicolor) and many non-red Spinosissimas.

Later, in the search for hardy bicolors, I crossed Hansa diploid with Hazeldean pollen and again got a large number of dwarfs all unhappy plants in one respect or another. However, some of the offspring were satisfactory. I got a dozen or so single bicolors of rather pale tone, and two rather charming semi-double bicolors, leaning toward red, and the other intermediate. I also got a true yellow. This told me that Hansa is heterozygous for its red-violet color just as all the yellow roses I have mentioned are heterozygous for yellow. Nevertheless, it was quite a surprise to find a yellow descendant of a rose so noted for strong red color. The best of these Hansa-Hazeldean crosses has been named Musician, and is a strongly colored bicolor, though with the red dominating the yellow. All of these Hansa-Hazeldean descendants fade quickly, and, therefore, if we would produce a satisfactory non-fading bicolor, a gene for sunworthiness must be garnered in from some other source.


In this cross I also noted another curious phenomenon. These bicolors contain genes for all three of the primary colors, red, yellow and blue, and so have an unmistakable grey tone in some cases. Therefore, if one wants an attractive hardy bicolor (rather than tricolor), he must cross his yellow rose with a clear pink rose and not with one of "rose" color-tone—or with a clear red, and not with a violet or crimson. I do not know if any pink rose exists which is completely free of all genes for blue, but certainly some come near enough to it to be regarded as satisfactory for all practical purposes. One of these is Rosa nitida, and another, perhaps, is Rosa rubrifolia.

All such experiments, seeking a clarification of the "color gene" picture, could undoubtedly be carried on with tender varieties more easily than with hardy ones, since, in the hardy group the sterilities and incompatibilities resulting from the mixture of R. spinosissima or R. foetida with other races of roses have yet to be overcome. In the tender group of roses these problems seem to have been overcome long ago in the origination and perfection of the world-famous Pernetianas.

CybeRose© note: We now know that carotene, the yellow pigment, is degraded and oxidized to produce the stronger perfumes found in roses. Thus, it is not necessary to assume that 'Harison's Yellow' is heterozygous for yellow. Rather, Wright's hybrids of R. macounii x 'Harison's Yellow' inherited the "gene for perfume" from the seed parent, which allowed all the yellow pigment inherited from the pollen parent to be converted to scent.

There is no blue pigment in roses. Mauve to violet tones occur when red cyanin is complexed with apigenin and/or gallotannin. Metal ions are probably involved. Iron has been suggested as a possibility.

Clear pink coloring is most easily accomplished by selecting for increased concentration of peonin, which does not "blue" in the presence of co-pigments like the more common cyanin. Saturated peonin is responsible for the violet-red of 'Hansa,' even without co-pigment.

Strange to say, 'Harison's Yellow' does carry "genes for red", possibly the same peonin found in R. foetida bicolor. The red pigment is usually suppressed (rather than absent), though it is sometimes expressed as in the picture at right. This plant always produced pure yellow flowers before and after this picture was taken.