American Rose Annual 1960
Inheritance of Magenta Red Color in Roses
Dr. Walter E. Lammerts

Head of Germain's Horticultural Research Division
Livermore, Calif.

One of the most important characteristics needed in an ideal red rose is a non-fading vivid color that does not "blue" with age. Unfortunately, from the genetic point of view, this is one of the most difficult combination of factors to obtain.

As mentioned in my article in The American Rose Annual for 1945, "The Scientific Basis of Rose Breeding," modern rose varieties, at least those used in modern rose breeding, are tetraploid. This means that they have four sets of the basic number of chromosomes characteristic of wild rose species i.e. four times seven, or 28 chromosomes. In most varieties, as a result of this, variable numbers of associations of four chromosomes occur. These are technically called quadrivalents. Thus Queen Elizabeth has from three to four even more quadrivalents, as determined by studies of pollen mother cells.

Secondly, the magenta type of coloration is the dominant one in roses characteristic of most of the naturally occurring species. Accordingly, in order to get clear non-fading red varieties, it is necessary to have all four of the chromosomes carrying this factor in the recessive condition. Using the terminology of Mehlquist, let us then represent the factor for magenta red by M and that for non-fading red by m. What should we then expect when we cross a variety such as Chrysler Imperial with Queen Elizabeth?

Careful study of progenies resulting from this type of cross indicates that many red roses which are at present very popular, and which are a vivid red under most climatic conditions do have at least one, and usually two of the dominant M factors. It is for this reason that roses such as Mirandy, Chrysler Imperial, New Yorker, and others fade out to magenta red and under cool weather conditions tend to be magenta red in color.

Colors Expected

The pattern of behavior of chromosomes which form quadrivalents has been rather fully explained in the discussion of the inheritance of glossy versus dull foliage in my article referred to above. Were it not for this sort of association, one would get a very large number of non-fading reds by crossing varieties such as Chrysler Imperial to Queen Elizabeth. Instead, only a small percentage are realized in each of the various progenies so far grown. This may best be explained by assuming that Chrysler Imperial carries two of the dominant magenta factors, or is MMmm. [In his 1964 article on the inheritance of pelargonidin he assumes that Chrysler Imperial is MMMm.] By classifying the various plants in a population obtained from Queen Elizabeth x Chrysler Imperial, using the British Horticultural Color Council Chart, one can readily group the colors obtained into, (1) dark purples, (2) shades which are either Tyrian purple or very similar to it, (3) shades which classify very similar to Solferino purple, (4) a large number which are a vivid rose-red similar to Chrysler Imperial, and (5) a few which are non-fading and are either current red, cardinal red, Turkey red, blood red, or Orient red. Crosses of Queen Elizabeth with Independence gave a small percentage of Solferino purple plants; whereas, Dean Collins x Independence gave no Solferino purples. Thus, the percentage of Solferino purples in one group of 260 plants of Queen Elizabeth x Independence was 2.7% or 7 plants, which conforms fairly close to expectation of 3.57% on the assumption that Independence is mmmm and Queen Elizabeth Mmmm.

On the basis of the above facts, we would expect the following behavior when Queen Elizabeth is crossed with Chrysler Imperial: (see Table I).

TABLE I

Gametes from
Chrysler Imperial
Gametes from Queen Elizabeth
1 MM 12 Mm 15 mm
6 MM 6MM
MM

Dark Purple

72Mm
MM

Tyrian Purple

90mm
MM

Solferino Purple

16 Mm 16 MM
Mm

Tyrian Purple

192 Mm
Mm

Solferino Purple

240 mm
Mm

Rose Red

6 mm 6 MM
mm

Solferino Purple

72Mm
mm

Rose Red

90 mm
mm

non-fading Currant Red

6 Dark Purple
88 Lighter Tyrian Purple
288 Solferino Purples and Lavenders
312 Rose Reds
90 Currant Reds
-----------------------------
784 Total plants

The above table is based on the assumption that the magenta factor is carried in one of the chromosomes which characteristically form quadrivalents and that as mentioned above, chromosome division occurs before the reduction division; so that instead of four chromosome strands segregating, there are actually eight. Each factor is thus represented eight times instead of the original four.

With such an association of eight chromosome strands or chromatids, any single chromosome may be segregated with any other and as a result we have a total of 28 gametes from Chrysler Imperial, 6 of which are MM, 16 Mm, and 6 mm. Also from Queen Elizabeth, we have a total of 28 gametes, one of which is MM, 12 Mm, and 15 mm. The random union of these gametes give the disribution shown in the above table. Analysis of the populations resulting form crossing Chrysler Imperial and 41093/161 with Queen Elizabeth are shown below: (see Table II)

TABLE II

Varieties Crossed

.76%

MM
MM

Pansy Purple
garnet lake

11.5%

MM
Mm

Tyrian
Purple

36.7%

MM
mm

Solferino
Purple

39.7%

Mm
mm

Rose-Red
& Pinks

11.4%

mm
mm

non-fading
Currant Reds

Total

Actual Kinds

Theoretical Expectation

Queen Elizabeth x Chrysler Imperial

1

.6156

*

10

9.31

*

33

29.73

*

27

32.15

*

10

9.2

*

81

81

*

Actual Kinds

Theoretical Expectation

41093/161

[Captain Thomas (yellow) x Crimson Glory]
pollinated by 46006/44 (semi non-fading)

1

1.4 *

*

33

20.5 *

*

63

65.3

*

*

65

70.6

*

*

16

20.2 *

*

178

178 *

*

41093/161 is a red variety very similar in color to Chrysler Imperial which has been very valuable in our breeding program because it is practically immune to mildew. 46006/44 is a semi non-fading red very similar to Starfire in behavior. It would seem then, from the very close fit of the actual color groups observed in these populations with the expected ratio, that red varieties of a color similar to Chrysler Imperial have two of the dominant magenta factors as assumed.

Of further interest are the results obtained by crossing the mildew immune variety 41093/161 with Golden Scepter and Yellow Perfection. In this case the various reds are classified into magenta reds and non-fading reds as shown by the following: (see Table III)

TABLE III

Magenta
Reds
Non-fading
Reds
Total
Reds
Light
Yellow
Deep
Yellow

Total

41093/161 x Golden Scepter

MM mm
mm mm

Actual Kinds:

32 14 46 14 4 64
Expected Ratio Among Reds:

6MM -- 16Mm -- 6mm
mm ----- mm ---- mm

Among yellows (see below)

39.3 11 50.3 10.3 3.4 64
41093/161 x Yellow Perfection

MM mm
mm mm

Actual Kinds:

46 9 55 8 -- 63

Expected Ratio Among Reds:

6MM -- 16Mm -- 6mm
mm ----- mm ---- mm

Among yellows (see below)

38.9 10.6 49.5 10.1 3.4 63
6/28 of gametes are free of basic color factor A so can be yellow, i.e., 6/28 of 64 = 13.71 compared to 18 in
Golden Scepter population; 6/28 of 63 = 13.50 yellows compared to 8 in Yellow Perfection population.

We again get a very close fit to expectation which further proves that varieties such as 41093/161 carry the two dominant magenta factors since, of course, Golden Scepter and Yellow Perfection do not have either the dominant magenta factor or the basic anthocyanin factor A. As may be seen, the percentage of yellow seedlings which may be expected from crossing varieties such as 41093/161 (Crimson Glory x Captain Thomas) to yellow varieties is approximately one-quarter. This would seem to indicate that either the basic factor for anthocyanin is in one of the chromosomes which do not form quadivalents, or at least is segregating independently of the factor for magenta coloration.

Actually in such a small population, it is impossible to distinguish between the sort of ratio one would get on the assumption of chromatid segregation in quadrivalents as compared with segregation in bivalents. Thus we would be comparing 6/28th with 7/28ths and would of course have to grow vary large populations to determine which of the two behaviors actually is occuring.

It is further interesting to note that deep yellow is recessive to light yellow. This, of course, I reported before in the 1945 paper. The problem of developing deep yellow varieties is accordingly even more complicated than that of developing non-fading reds, since we must not only get rid of the magenta factor, but the basic dominant color factor A for anthocyanin and the dominant factor Y for light yellow.

As may readily be seen from the above discussion of the facts of color inheritance so far found, the breeding of both non-fading red roses and yellow roses is no easy task. Basically, we need breeding lines which are true breeding for the non-magenta m factor combined with such desirable traits as upright, compact habit of growth, abundance of wide, leathery, glossy foliage, and a high degree of resistance to mildew, blackspot, and rust. Several such lines have finally been established, but of course, much more refinement is needed as regards such all-important qualities as bud and flower form, substance, and fragrance, before any of thses can be introduced.

The problem of getting rid of even such a single non-desirable factor as the magenta red one discussed above, and combining non-fading red with large, glossy, mildew-proof foliage and fragrance, is a very time consuming one in a plant sohighly variable as the rose. Fortunately this phase of the breeding work so necessary in the development of really non-fading roses is now complete in that we have several breeding lines which are very vigorous and combine the above desirable qualities.

Other commercial breeders are making substantial progress in the development of lines breeding true for deep yellow, and I am sure, therefore, that in the next decade, garden rose enthusiasts can look forward to varieties which are really basically improved and not merely varieties with some minor feature of distinction.