American Rose Annual (1954) p. 73-77
Notes on the Breeding Behavior of Rosa Roxburghii and Rosa Multibracteata
Dr. H. D. Wulff

Botanical Institute of the University of Kiel, Germany

IN THE recent literature I found some erroneous statements on the origin of the hybrid polyantha rose 'Floradora' which need correction. In Modern Roses III (1947) and IV (1952) this rose is cited as a hybrid between the well-known polyantha rose 'Baby Château' and Rosa multibracteata. W. Kordes (Rosenjahrbuch vol. 1, 1950, p. 52) as well as R. E. Shepherd (American Rose Annual vol. 38, 1953, p. 153) also mention this origin. During the last three years I had the great pleasure of a close co-operation with the late German rose breeder, Mr. Mathias Tantau sen., who died on June 26th, 1953, nearly 71 years old. Our deepest thanks are due to Mr. Tantau for a considerable number of beautiful garden roses created by him, amongst these 'Floradora'. Mr. Tantau personally informed me that three of his best hybrid polyantha roses, 'Floradora', 'Käthe Duvigneau', and 'Cinnabar' (in German, Tantau's Triumph) had originated from a cross between 'Baby Château' as mother and Rosa Roxburghii (not R. multibracteata!) as father plant.

There cannot be any doubt that 'Baby Château' is a tetraploid rose (with 28 somatic chromosomes or 14 chromosomes in its sex cells) as most of our modern garden roses are, and one of my pupils will give fuller information on its cytology in due time. Rosa Roxburghii, on the other hand, is known to be a diploid species with half as many chromosomes, i.e., seven in its sex cells. Judging from these chromosome numbers it was expected that 'Floradora', Käthe Duvigneau, and 'Cinnabar' would be triploid roses with 21 chromosomes, 14 of them originating from the female and seven from the male parent. It is a very remarkable fact that my cytological investigations proved all of them to be tetraploids. Triploid plants, as a rule, are totally sterile—but the mentioned three roses set seed abundantly, as tetraploids generally do. It was possible, therefore, to make use of them in further crosses from which, among others, the hybrid polyantha rose 'Improved Cinnabar' (in German, Verbesserte Tantau's Triumph) arose, this latter being the hybrid (Cinnabar x Käthe Duvigneau) x Cinnabar.

Mr. Tantau also made an attempt to cross hybridize tea roses with Rosa Roxburghii. His hybrid "No. 46534," for instance, originated from a crossing of 'Cerise' with the last mentioned species as male partner. As to the chromosome number, we meet with the same phenomenon as referred to above. 'Cerise' possesses 28 somatic chromosomes, Rosa Roxburghii 14, and the hybrid 46534 also 28 instead of the expected 21 chromosomes!

At the present moment I am unable to give a precise answer to the question by what mechanism the increase of the chromosome number in the hybrids might have been achieved. Three explanations are possible: (1) R. Roxburghii is existing not only in diploid (14) but also in tetraploid (28) races, (2) R. Roxburghii, if diploid, is producing unreduced pollen grains in a rather high amount, (3) the chromosomes of R. Roxburghii are undergoing a supernumerary splitting shortly before or after fertilization. Further investigations will be necessary to elucidate this problem.

I may refer here briefly to a similar cytological phenomenon which was detected in the Lambertiana rose 'Eva' by H. D. Wulff and L. Heldt (Zuchter vol. 23, 1953, p. 87-93). 'Eva' was produced by Mr. W. Kordes from a cross between the diploid Lambertiana rose 'Robin Hood' and the tetraploid hybrid tea rose 'J. C. Thornton'. It has also not become triploid but is a tetraploid rose like the male parent. A precise explanation for the increase of the chromosome number could not be given in this case either, until now.

There is another remarkable fact to note. Without going into details I may state here that the three roses 'Floradora', 'Käthe Duvigneau', and 'Cinnabar', as well as the hybrid 46534, did not show any traces of the male parent Rosa Roxburghii in their morphology. The first three roses are true hybrid polyanthas, the latter is a true hybrid tea, indicating thus that the genes which are responsible for the respective characters of growth habit and for many characters of shape and size of flowers, fruits, leaves and spines are dominant to the allelic genes of R. Roxburghii. Only anatomical studies revealed a certain similarity and relationship to the latter species.

An orange-red tone in the color of the petals is very pronounced in 'Floradora' and 'Cinnabar', less in 'Käthe Duvigneau'. We shall see afterwards that this orange-red color, which is often met with in the progenies of 'Baby Château', is also of importance for the R. multibracteata crosses of Mr. Tantau sen.

To sum up, 'Floradora' possesses genes of R. Roxburghii but not of R. multibracteata. And as far as I know, there is not yet existing any garden rose which contains genes of the latter species. I only know of two breeding experiments of the German nursery men, Mr. Wilhelm Kordes and Mr. Mathias Tantau sen, who both tried to introduce the genes of Rosa multibracteata into our garden roses. Mr. Kordes succeeded in crossing his 'Baby Château' with Rosa multibracteata, but the double flowers of the hybrid did not open at all and produced no or only a few anthers. Some more details on this hybrid will be published by one of my pupils. Mr. Tantau sen., on the other hand, chose a hybrid tea rose "X" (the true name of which cannot be given here) for the cross with Rosa multibracteata and, in 1938, got a hybrid "A" which was similar to the male parent in most characters, but the flower was double and beautifully red-coloured. Another hybrid tea rose "Y" was pollinated with "A", and the arising next generation (hybrid B) highly approached a hybrid tea rose in its appearance.

The whole pedigree becomes branched and highly complicated at this point, because the single hybrids were now used both as female and male parents and often gave rise to more than one new hybrid which deserved attention and further breeding. Only one line, therefore, may be traced here, beginning with the cross "B" x "120" (120 being the hybrid Arabella x Heros) which led to the hybrid C, a rose distinguished by vigorous growth, healthy foliage and very large, fragrant flowers with about forty red petals. One plant (D) of the next hybrid generation, which arose from the cross C x 4167, is very important for two reasons. Firstly, the rose 4167, used as male parent, is a hybrid between 'Cinnabar' and 'Käthe Duvigneau', and with it genes of Rosa Roxburghii have been brought into this R. multibracteata progeny. Secondly, the hybrid D is distinguished by flowers with about fifteen petals of an intense orange-red color. I may refer here to my above remarks and repeat that this orange color probably depends on the action of genes derived from 'Baby Château' which is grandmother to the hybrid D. This latter was back-crossed with 4167. The arising hybrid E was of vigorous growth and abundantly produced double flowers of a bright orange color. Unfortunately the flowers were not quite upright, but hung over a little bit, so that further improvement by hybridization was necessary. This was done by the cross E x 4503, the latter being the hybrid R.M.S. Queen Mary x 4167.

The resulting last rose of this line (52506), the fruit of fifteen years' breeding work, was regarded by Mr. Tantau sen. himself as his best hybrid tea rose. Under greenhouse conditions it grows straight upright, reaching a height of about thirty-one inches. The foliage is resistant to fungus diseases and the double flowers are attractively orange-colored and remain in good condition for nearly two weeks. They consist of about thirty petals which are surrounded by five sepals showing deep incisions and sepalet formation in a compound manner. They do not curve downward when the flower is opening but spread out sideward.

The R. multibracteata-hybrids mentioned here and others which I investigated during the last two years are of good or high fertility and offer a rather normal cytological behavior, for Rosa multibracteata and all ornamental roses crossed with it, are tetraploids. During the reduction division in the pollen mother cells the hybrids of R. multibracteata as well as those of R. Roxburghii show, as a rule, the normal formation of 14 bivalents from the 28 univalent somatic chromosomes. Even in the hybrids D and E, which contain chromosomes of both these wild species, the chromosome pairing during the reduction division occurs normally in most of the pollen mother cells. We may conclude, therefore, that the basic chromosome sets of all wild rose species are homologous to a very high degree. Furthermore, the preponderance of pollen mother cells with 14 bivalents and the nearly total absence of multivalents seem to speak in favor of the view that bivalent formation in roses is controlled by genes.

Some abnormalities are, nevertheless, to be observed in a very low percentage of the pollen mother cells. They are of the same type which I have described for the reduction division of Rosa Kordesii in the preceding volume of this annual (vol. 38, 1953, p. 111-122) and consist of the presence of univalents and multi-valent rings or chains. They indicate that some of the chromosomes have undergone structural changes by so-called translocations.

'Floradora', 'Cinnabar', 'Käthe Duvigneau' and the related 'Improved Cinnabar' represent successful attempts of introducing genes of Rosa Roxburghii into the hybrid polyantha roses. These four roses have flowers of very attractive colors and shapes, are of good hardiness and—last but not least—are of exceedingly high resistance to fungus diseases. Improvement of our hybrid tea roses is far more necessary than of the hybrid polyanthas. It has been demonstrated above that the introduction of R. multibracteata and R. Roxburghii genes into the hybrid teas is possible and leads to astonishing results under greenhouse conditions. It may be—and future observations will have to prove it—that Mr. Tantau sen. has opened up a new way for better hybrid tea roses which will also compete for a place in the amateur's garden.

CybeRose Note: Another possibility is that Tantau's three seedlings from 'Baby Château' x R. roxburghii were partial hybrids. Most of the R. roxburghii chromosomes were eliminated, and those from the seed parent were duplicated. Only small bits of R. roxburghii chromosomes would remain. This phenomenon is known among various grains, Helianthus, Solanum and other genera.

I take it that rose "X" is 'Crimson Glory', and its hybrid with R. multibracteata is 'Cerise Bouquet'.

Alpine Glow: Cinnabar x (Cinnabar x Käthe Duvigneau)

Cerise: Crimson Glory x Sterling (Tantau, '45)

Sterling: Mme. Butterfly x seedling (E. G. Hill Co., '33)

4167: cross of Cinnabar and Käthe Duvigneau

120: Arabella x Heros

4503: R.M.S. Queen Mary x 4167