American Rose Annual, pp. 120-124 (1961)
Observations On the Inheritance of Yellow Pigment in Roses
Dennison Morey
Livermore, Calif.

THE INHERITANCE of pigments in living organisms is most marked by the wide diversity and seeming ingenuity of the many systems which have been developed in different organisms through the processes of evolution. Yet, though there is great diversity in the details of pigment inheritance, certain basic similarities are often felt to exist and have been proven to exist in many instances between very different kinds of organisms. For example, in both plants and animals, color inheritance is most often based upon various modifications of a system involving a set of four genes, each of which affect, in significant measure, the ultimate color.

This basic set of four factors consists first of a gene for the presence or absence of any and all pigment and is the fundamental controlling factor for albinism and the modifications of albinism such as the Siamese pattern in cats and the Himalayan pattern rabbit. Next there are the factors for the kind of color that will be produced; e.g. whether the color of a horse is to be black or brown. It is at this point that we run into complications because colored hair usually has two colors—an inner color which is usually red or yellow, and an outer color which is black or brown. Either of these may be lacking or present in numerous combinations giving rise to the wide variety of colors which are evident in all hirsute organisms. In plants we also have a two gene, two color system, red and yellow, which may occur in the same flower and which produces the same sort of blending effects that one sees in animals.

In addition to the factors for actual qualitative color there are ordinarily to be found in any organism, modifying factors for the intensification of color or conversely for the dilution of whatever color is present, or for causing qualitative color variations of the same basic pigment.

As has been pointed out, this pattern seems to be basic to both plants and animals and to all higher plants and animals which have the ability to produce pigments. As a consequence, there is often a strong tendency to try to make all examples fit this or closely related patterns without giving adequate credit for possible ingenuity and the possible creation of exceptions to the underlying universal pattern on the part of these always unpredictable living laboratories.

One of the basic patterns in plants, and one often considered typical, is that studied in connection with the color of carnations. This study by Mehlquist and Geismann(1) showed that carnations have a factor for suppression of pigment formation (though apparently not for all pigment, only the yellow being affected, the anthocyanin or red-blue pigment being unaffected by the suppressor gene); that they have separate factors for yellow and red colors which may occur separately or in combination; and that the development of red-blue colors depends upon two different modifiers or secondary color factors. In addition there appears to be an intensifier factor for the red-blue pigment. Thus in essential aspect, there is very little difference between cavies and carnations as far as pigmentation is concerned. The designation for the situation in carnations is YIASRM. (The "Y" represents the gene for yellow color, "I" is the suppressor of yellow, "A" is the gene for red-blue color, "S" is its intensifier, "R" is the gene which determines the relative redness or blueness of the red-blue pigment and "M," which is of similar action to "R.")

Inasmuch as modern roses are characterized by both red and yellow pigments, it would certainly be convenient if the work on other plants such as that on carnations(1) and dahlias(2) could be used as a basis for evaluation of pigment inheritance problems in roses. That such is or is not possible is not the easiest thing to prove since roses are for the most part tetraploid and very difficult to analyze even when adequate populations have been produced, a feat of no mean proportions itself.

It is because detailed analyses of hybrid garden rose populations are economically impractical that the convenience of drawing safe inferences from, and the confident planning of crosses upon data obtained in other plants, assumes great practical importance. The rose breeder could expect to make far greater, faster, and more economical progress if he had some means of knowing what the probable genetical situation was in any given rose variety by looking at it and a limited number of its offspring. If roses had the YIASRM pattern, such a procedure should be possible and highly productive.

Much study and observation indicates that such a basic pattern may very well apply to roses. However, certain very real deviations are quite apparent and are of fundamental importance. While garden roses are for the most part tetraploid, they seem in most instances to segregate as diploids. However, one must realize that any characteristic which segregates as a single factor difference in diploid forms, will segregate as a two factor difference in the tetraploid and that even a little multivalent formation will confuse things very badly.

I have had at hand for several years the results of a study of the segregation of yellow color in tetraploid garden roses. The results are summarized in Table 1.


Color Pale
Frequency Observed 3 15 30 24 7
Frequency Expected 5 20 30 20 5

X2 = 3.8 Very highly significant (99% of all random samples would be expected to vary more.)

This table shows the data fit a ratio that results from a system of duplicate additive genes with partial dominance. I think this is to be fully expected since we know there must be a total of four loci on the chromosomes for yellow color. We also strongly suspect that there are several different kinds of genes for yellow color. I have long suspected that there were two. One of these I felt was the form which produced true yellow and the other, cream or ivory. I had also felt that the deep butter yellow of Rose d'Or and Lydia could only be obtained from having all four yellow genes in the dominant 100% additive condition in a garden rose and that in diploid roses the most intense color that one could expect was that seen in R. hugonis or Harison's Yellow.

With this hypothesis my yellow breeding program has produced some very gratifying results both from the theoretical and practical point of view. However, even though the theory was productive, it has proven to be somewhat in error. I discovered these errors by means of studying diploid roses.

The miniature roses are diploid for the most part although the yellow varieties are for the most part triploid or aneuploid. They do, however, offer a practical means of checking color inheritance since there are presumably only two rather than four loci present for each characteristic controlled by a single gene system. They offer the best and only practical means of determining the number of different forms which a color gene may have. While it doesn't concern us here, I am now confident that the red gene for instance comes in at least five forms some of which may be due to modifiers such as "R" and "M" in carnations but three of which are certainly basic. The miniature roses are now available in nearly all of the colors to be found in their larger cousins. Moreover, my newest kinds, many of which are yet to be introduced, are all fully dwarf diploids. These afford an excellent means of now testing the YIASRM theory.

So far I have only had the opportunity to use them to check certain important details of color inheritance such as the allelar situation. The information gained is very interesting especially as far as yellow pigment is concerned, since I now know that my original theories about yellow color are far too simple, even though duplicate additive partial dominance may seem quite involved enough without complications. In addition to yellow and ivory factors, there is either a third form of the yellow gene or else there is an intensifier gene for yellow. The data are summarized in Table II.


  Pale Light Med. Dark Intense
Diploid   X X X  
Tetraploid X X X X X

The same butter yellow color can be produced by a diploid plant with only two genes as is found in tetraploid roses with twice as many genes. A higher level of pigment saturation does seem to occur in the tetraploids, however.

The only conclusion which I would care to draw at this time is that either the diploid deep colored yellow roses have an intensifier gene acting and the tetraploids do not, or else one is forced to conclude, at least for the moment, that there are three forms of the gene for yellow color. Were the intensifier present in all roses, then the inheritance in the yellow garden rose analysed would have segregated differently. From a great deal of breeding work we know that pigment suppressors occur in roses and there is every reason to believe that intensifiers as well as modifiers also exist. Moreover, whether or not the yellow and red pigment systems interact has not been determined for roses, but it seems entirely possible that they do.(3) Thus one cannot safely rule out the possibility that the intense yellow in the diploid seedlings discussed may not be due to an intensifier. This possibility is being checked.

In roses lacking red pigment systems, the inheritance of yellow pigment appears to be a relatively simple system based upon the duplicate additive segregation of two genes which have two and possibly three qualitative values or forms but a system which is complicated in many instances by the presence of pigment intensifiers, suppressors, and probably several modifier genes affecting the color quality. The deepest yellow color seems to be obtained only when the suppressor and modifier genes are in the recessive condition.


  1. Mehlquist, G. A. L. and Geismann, T. A. 1947. Inheritance in the Carnation, Dianthus caryophyllus. III Inheritance of flower colour. Ann. Mo. Bot. Gard. 34:39-72.
  2. Lawrence, W. J. C. 1931. The Genetics and Cytology of Dahlia variabilis. Journ. of Genetics. 24:257-30.
  3. Lawrence, W. J. C. and Scott-Moncrieff, R. 1935. The Genetics and Chemistry of Flower Color in Dahlia; a new theory of Specific pigmentation. Journ. Genet. 30:155-226.