THE GARDENERS' CHRONICLE. 84: 35-36 (July 14, 1928)

GENETICS OF THE ROSE.*
*A paper contributed to the international Rose Conference by Major C. C. Hurst.

SINCE the War I have devoted myself entirely to a study of the genetics of the Rose. A comprehensive collection of the known species, sub-species and hybrids has been got together in the Cambridge Botanic Garden from various sources, including wild species collected by my wife and myself in England and in five Cantons of Switzerland, and plants raised from seeds sent to me by correspondents and travellers in North America, Mexico, Turkestan, Siberia, China and Japan. I am also much indebted to the Directors of the Botanic Gardens at Kew, Oxford, Cambridge and Basel; to Colonel Gravereaux, of La Roseraie de l'Hay, Paris; and to Dr. Mary Carew-Hunt (who kindly allowed me to select eighty specimens from the collection of the late Canon Carew-Hunt at Albury, Oxford), for free access to their collections in obtaining my material. This material has been analysed in three different ways, in accordance with the methods of three different sciences; Taxonomy, Genetics and Cytology.

First.—About one hundred characters of each species, sub-species and variety have been examined and tabulated.

Second.—Numerous experimental crosses have been made between various species, sub-species and varieties, and the results recorded and tabulated. Many known hybrids have also been analysed.

Third.—The chromosomes of 674 species, sub-species, varieties and hybrids have been examined and counted under high-powered microscopes, in various stages in both body-cells and germ-cells.

The results of these combined experiments and researches have proved to be of considerable importance, and in many cases, surprising and quite unexpected results have been obtained. I fear that several large volumes will be required to record the results already secured, and so far as one can see, we are as yet merely on the threshold of important results to come.

Since, on this occasion, I have been allotted twenty minutes, I propose to devote ten minutes to giving a very brief account of the results obtained, and the remaining ten minutes to a consideration of the possibilities of applying these purely scientific results to the betterment of our garden Roses.

The most, important results have come from the counting of the chromosomes. A Rose plant, like other plants and animals, is made up of millions of minute cells which form the various tissues. Each cell contains a round, central body, known as the nucleus. Each nucleus contains a number of microscopical rod-like bodies which stain rapidly with chemical dyes, and for that reason are called chromosomes. These chromosomes carry the Mendelian factors or genes, which differentiate the various characters of the Rose. In the body-cells of regular species, the chromosomes consist of pairs, one of which came from the male parent and the other from the female parent. In the male and female germ-cells the chromosomes are single, so that a pollen-nucleus and an egg-nucleus, carry one-half the number of chromosomes that are contained in the body cells. When fertilisation takes place the two singles come together and make a pair of chromosomes. As an illustration of how this works in a given ease, we will take two well-known Roses and cross them together. The female parent is Rosa multiflora from Japan, with tall, summer-flowering stems and single, white flowers. The male parent is Rosa chinensis from China, with dwarf, perpetual-flowering stems and semi-double, pink flowers. Each parent has seven pairs of chromosomes in its body-cells and seven single chromosomes in its germ-cells. From experiments we know that the Mendelian gene for tall summer-flowering stems is located in one chromosome of the seven present in the egg-nucleus of R. multiflora, while the gene for single flowers is located in a second chromosome and the gene for white flowers is located in a third chromosome.

Similarly, the gene for dwarf perpetual-flowering stems is located in one chromosome of the seven present in the pollen-nucleus of R. chinensis, while the gene for double flowers is located in a second chromosome, and the gene for pink flowers is located in a third chromosome. After fertilisation, the cells of the hybrid contain seven pairs of chromosomes, seven from each parent. The chromosome carrying the gene for tall, summer-flowering from R. multiflora pairs off with the chromosome carrying the gene for dwarf perpetual-flowering from R. chinensis, as do the two chromosomes carrying genes for single and double flowers, and the two chromosomes carrying genes for white and pink flowers. The resulting hybrid produces tall, summer-flowering stems with semi-double pink flowers, since these characters are dominant, while dwarf perpetual-flowering stems and single white flowers are recessive. When the germ-cells of the hybrid are formed, the seven pairs of chromosomes are reduced to seven singles. Since this reduction is a random one, the hybrid produces eight kinds of pollen-cells and eight kinds of egg-cells. Consequently, when the hybrid is selfed, in accordance with the Second Law of Mendel, we get, on the average, once in sixty-four times a dwarf perpetual-flowering double white Rose, with the mixed characters of R. multiflora and R. chinensis. Such was the origin of the Polyantha Pompon "Paquerette," raised by M. Guillot in 1873, which introduced a new race of Roses to our gardens.

ćA more complete account appears in Report of the
5th Genetics Congress, Berlin, 1927, Vol. 2 (1928).

The chromosomes in Roses, as in other plants and animals, are relatively constant in size, shape and number for any particular individual. In the five genera of the Rose Tribe there are minor differences in size and shape of the chromosomes, but in Rosa proper all the chromosomes are relatively the same in size and shape, though they differ remarkably in number. Six different types have been found with fourteen, twenty-one, twenty-eight, thirty-five, forty-two and fifty-six chromosomes in the body-cells. The interesting and important point is that all these numbers are multiples of seven. In the formation of the germ-cells and at certain other times, the chromosomes may be seen working in sets of seven, or septets as we call them. To cut a long story short†, our genetic experiments at Cambridge have demonstrated that in Rosa proper there are five distinct septets, or sets of seven chromosomes, which we distinguish as A, B, C, D, and E septets. Each septet of chromosomes carries a different set of genes representing at least one hundred specific, sub-specific and varietal characters. From this it follows that there are five fundamental species in the genus Rosa of Linnaeus, and that all the other species of this genus consist of various combinations of the chromosomes and characters of these five fundamental species. The five fundamental species are known as diploids because they contain two septets of chromosomes in their body cells, other Roses are known as polyploids because they contain more than two septets. Triploids contain three septets, tetraploids four, pentaploids five, hexaploids six, and octoploids eight septets of chromosomes in their body-cells.

The first fundamental diploid species which carries a pair of A septets of chromosomes is Rosa sempervirens of Linnaeus, and its subspecies include R. arvensis, R. moschata, R. phoenicea, R. abyssinica, R. Pissartii, R. Brunoni, R. Leschenaultiana, R. longicuspis (with its variety lucens), R. gigantea, with its offspring R. odorata (the Tea Rose), R. Soulieana, R. Helenae, R. Rubus, R. Gentiliana, R. laevigata (with its offspring Sinica Anemone) R. Banksiae, R. microcarpa, R. chinensis (with its offspring the pink and crimson Chinas), R. anemonaeflora, R cathayensis (with its offspring Crimson Rambler), R. multiflora, R. Wichuraiana, R. Luciae, R. Watsoniana, R. rubifolia, R. setigera and several new sub-species from China found at Kew and Cambridge in Mr. R. Cory's seedlings.

The second fundamental diploid species which carries a pair of B septets of chromosomes is Rosa sericea of Lindley, and its sub-species include R. cabulica (usually mislabelled R. Beggeriana in gardens), R. Ecae, R. Webbiana, R. omeiensis, R. sertata, R. Willmottiae, R. Hugonis, R. xanthina (of American gardens, not the one in British gardens, which is a tetraploid R. ochroleuca with B and D septets), R. gymnocarpa and a new sub-species found at Cambridge.

The third fundamental diploid species which carries a pair of C septets of chromosomes is Rosa rugosa of Thunberg, and its sub-species include R. coruscans, R. nipponensis (not R. acicularis nipponensis of gardens, which is a tetraploid with D and E septets), R. nitida and two new sub-species found at Kew and Albury.

The fourth fundamental diploid species which carries a pair of D septets of chromosomes is Rosa Carolina of Linnaeus (of 1753, not his species of 1762, which is, in part, a tetraploid with A and D septets), and its sub-species include R. Cinnamomea of Linnaeus 1753 (not the species of 1762, which is a tetraploid with D and E septets). R. davurica, R. Marrettii, R. pisocarpa, R. Woodsii, R. Fendleri, R. foliolosa, R. blanda, and several new sub-species raised from seeds collected in North America.

The fifth fundamental diploid species is Rosa macrophylla of Lindley, and its sub-species include R. corymbulosa, R. Giraldii, R. elegantula, R. persetosa and two new sub-species from China found at Kew.

THE GARDENERS' CHRONICLE. 84: 54-55 (, 1928)

POLYPLOID ROSES

The other species of Rosa are polyploids with more than two septets of chromosomes, and these are all made up of various combinations of the chromosomes and characters of the five fundamental diploid species.

Regular tetraploid species with four septets are:—Rosa Huntii, a new species from China with a pair of A and a pair of B septets.

R. centifolia, the oldest Rose in cultivation with a pair of A and a pair of C septets. This species has numerous sub-species including R. damascena, R. gallica, R. provincialis, R. pumila, R. parvivolia, and R. pomponia (Rose de Meaux).

R. palustris, with a pair of A and a pair of D septets, with its sub-species R. corymbosa, R. Hudsoniana, R. humilis grandiflora (of gardens), and in part R. carolina of Linnaeus, 1762 (not 1753).

R. Davidii, with a pair of A and a pair of E septets, and its sub-species R. setipoda, R. rosea-Moyesii, R. Fargesii and R. crasseaculeata.

R. spinosissima, with a pair of B and a pair of C septets, and its sub-species R. myriacantha, R. Ripartii and R. hispida of Sims (not some R. hispida of gardens, which are hybrids).

R. pimpinellifolia, with a pair of B and a pair of D septets, and its sub-species R. hemisphaerica, R. lutea, R. altaica, R. lutescens, R. ochraleuca, R. grandiflora and R. Rapinii.

R. multibracteata, with a pair of B and a pair of E septets, and its sub-species R. reducta and R. bella.

R. virginiana (= lucida), with a pair of C and a pair of D septets, and its sub-species R. baltica, R. suffulta, R. Lunellii and R. saturata.

R. pendulina (= alpina), with a pair of D and a pair of E septets, and its sub-species R. pyrenaica, R. laxa of Retzius (Not of Lindley, or the garden R. laxa, which is a pentaploid), R. oxyodon, R. lagenaria and R. intercalaris.

Regular hexaploid species with six septets are comparatively rare;—R. Wilsonii, with a pair of A, B and C septets; R. manca of Greene, with a pair of A, B and D septets; R. Moyesii, with a pair of A, B and E septets; R. nutkana, with a pair of A, D and E septets; R. Bourgeauiana, with a pair of B, C and D septets; R. Engelmanni, with a pair of B, D and E septets; and R. Sayi, with a pair of C, D and E septets.

Regular octoploid species with eight septets are even more rare; R. Tackholmii, a new species aith a pair of A, B C and D septets; and R. acicularis with a pair of B, C, D and E septets. This is a circumpolar species.

Irregular tetraploid species with both paired and single septets are R. omissa, with a paired A and single C and D septets; R. recondita, with a paired C and single D and E septets; R. mollis, with paired D and single C and E septets; R. pomifera, with a paired E and single C and D septets; and R. rubrifolia, with a paired D and single A and E septets.

Irregular pentaploid species with both paired and single septets are:— R. agrestis, with a paired A and single B, C and D septets; R. canina, with a paired A and single B, D and E septets;

Irregular hexaploid species with paired and single septets are R. Pouzinii, with a paired A and single B, C, D and E septets; R. inodora, with a paired B and single A, C, D and E septets; R. Jundzillii, with a paired C and single A, B, D and E septets.

The experimental proof of these analyses rests entirely on genetical experiments, many of which have been carried out already. Time will not allow them to be shown here, but last summer I was able to demonstrate these experiments to the Genetical Society when they met at Cambridge.

From the scientific point of view our experiments with Roses have thrown a good deal of lint on the problem of species. The old and much asked question, What is a species? is now, so far as the Rose is considered, no longer a question of the personal opinion of doctors who differ, it is simply a matter of genetical experiments combined with the counting of the chromosomes in critical cases, while in ordinary cases the species may be determined by the use of a taxonomic table of the septet characters of the five fundamental species.

The point of real importance is that with modern methods of analysis we now have experimental proof that a species is a real entity made up of a chromosome complex of associated genes. The old systematists instinctively recognised this although in their day they were unable to demonstrate it experimentally.

By the same methods we are also able to determine definitely the sub-species within the species, and the varieties within the sub-species. This clears the way for a better understanding of the distribution of Roses in space and time, and in the end will no doubt help solve the ultimate problems of the evolution and origin of the species.

IMPROVEMENT OF GARDEN ROSES.

Now we come to consider how we may apply theswe new scientific facts to the improvement of our garden Roses. Perhaps the best way to do this will be rapidly to go through our garden Roses and see what they really are in their septet characters. The oldest cultivated Roses, R. centifolia, R. gallica and R. damascena, cultivated probably by the Greeks, Romans and Egyptians, are tetraploids with paired A and C septets. The old Chinas and Tea Roses are diploids with paired A septets, or triploid and tetraploid varieties with three or four A septets. The old Noisettes and the new Polyantha Pompons, are diploids with a pair of A septets. The boursault Rose is a diploid hybrid with single A and D septets. The old Bourbon is a triploid hybrid with a paired A septet and a single C septet. The old Hybrid Perpetuals are for the most part tetraploids with paired A and C septets. The old Hybrid Teas are tetraploids with three A septets and one C septet, while some of the modern H.T.'s are tetraploids with four A septets. The original Pernetiana seedling was a tetraploid with single septets of A, B, C and D, while later varieties have three septets of A and one B septet. The old natural hybrid Rosa x alba and its offspring "Maiden's Blush" is a hexaploid with paired A and C septets and single D and E septets. Most of the Penzance Hybrid Sweet Briars are hexaploids with paired A and C septets and single B and D septets, although "Lady Penzance" has paired B and D septets and single A and C septets, while "Catherine Seyton" is a pentaploid with a paired A septet and three single B, C and D septets.

From these analyses it is evident that in order to break now ground in Roses it is necessary to introduce the chromosomes and characters of the E septet into our present A, B, C and D Roses. This may be done by crossing with such species as Rosa Moyesii, a hexaploid from China with paired A, B and E septets, or R. nutkana, a hexaploid from North America with paired A, D and E septets.

A similar result can be obtained by crossing with R. rubrifolia a tetraploid from the Alps with a paired D septet and single septets of A and E. To get the desired result, however, care must be taken to use this species as the female parent, since the pollen cells only carry the single D septet which gives the characters of R. cinnamomea only, while the egg-cells carry the three septets A, D and E, which give the characters of R. rubrifolia. In the same way the hexaploid natural hybrid Rosa alba carries in its egg-cells the four septets A, C, D and E, while the pollen-cells carry only two septets, A and C. It is only by using R. alba as the female parent that one can get the R. alba characters, in using R. alba as the male parent one only gets the characters of R. damascena.

Hybrids between garden Roses and the above species are more likely to be fertile than direct hybrids with the pure E species Rosa macrophylla, which would give for the most part sterile triploids. Fertility in the second generation is necessary in the formation of a new race of Roses, since it is only in the second generation that we can get the recombinations of characters that we require and most of the desirable qualities in garden Roses are recessive in the first generation.

We now know definitely that fertility is largely bound up with the pairing of the chromosomes in complete septets. The chief desirable characters carried by the E septet are smooth, erect, cane-like stems, graceful foliage, large brilliant flowers and long, pendulous fruits of many colours. These characters would be a welcome stems, graceful foliage, large brilliant flowers and long, pendulous fruits of many colours. These characters would be a welcome addition to our garden Roses, and may be secured in the second generation of breeding, if the breeder has a bit of luck and follows it up. In attempting to add these new characters, however, great care must be taken not to lose the desirable characters of form, colour, fragrance and perpetual flowering that we already have, and here arises a real difficulty, which a knowledge of the septet characters will help us to overcome. For instance, the A septet provides us with the delicate translucent colours of the China and Tea Roses, while the combinations of the A and C septets gives us the deep velvety crimson of the Hybrid Perpetual, as well as the brilliant scarlets.

The B septet gives us the rich yellow of R. lutea which comes out in some of the Pernetianas. The C septet provides us with the delicate and refined old Rose fragrance of the Cabbage Rose, and the A septet gives us the musky odour of the Musk and Tea Rose, while the combination of both A and C septets produce the rich damask perfume of some of the old Hybrid Perpetuals. The B septet gives us some disagreeable odours, the D septet provides some spicy scents, while the E septet gives little fragrance. True perpetual flowering is peculiar to the A septet. It is unknown in a wild state, although I have observed two definite cases in wild seedlings raised at Cambridge. It first appeared as a recessive mutation in Chinese gardens in the sub-species R. chinensis. The gene for perpetual-flowering is either identical or very closely linked in the same chromosome with the gene for dwarf habit of growth. This would assist the breeder materially in creating a dwarf perpetual R. Moyesii.

There are also further but more remote possibilities in raising entirely new races of garden Roses by the use of four other distinct diploid species usually classed as Rosa; but which our septet analyses prove to be four distinct though related genera of the Rose tribe. One of those, Hesperhodos minutifolia, with its sub-species H. stellata and H. mirifica, so far refuses to cross with Rosa. This primitive genus from the Western United states is a desert plant with small leaflets resembling some of the fossil Roses found in the Miocene beds of Colorado, which are the oldest Roses known. It has large, striking flower and prickly fruits, and in the distance the plants is not unlike a Gooseberry Bush.

Two of the other genera have produced hybrids with Rosa, namely Ernestella bracteata from China and India, which, though rather tender, has large white autumnal involucrate flowers and hairy stems; it has been crossed with a yellow Tea Rose giving the beautiful "Mermaid" which, although bearing single flowers, is quite sterile at Burbage and Cambridge.

Platyrhodon microphylla, from China and Japan, the so-called Chestnut Rose with cup-shaped fruits covered with fleshy spines, seems more promising in its fertility since I have succeeded in raising the second generation of a cross with Rosa rugosa. The results, however, although extremely interesting from the scientific point of view, are not very promising horticulturally, since the grandparent R. rugosa has been reproduced in facsimile several times, while the others resemble the parent hybrid with strange mutational variations.

The remaining genus, Hulthemia persica, from the salty deserts of central Asia, is the Rose with simple leaves and no stipules, with Cistus-like flowers, yellow, blotched with red; it has been crossed with Ernestella, with which it gave the beautiful hybrid known as R. Hardii, which, however, is quite sterile at Burbage and Cambridge, and is difficult to keep alive out-of-doors. Since one of the parents comes from the hot plains of Bengal, this is not altogether surprising, and the other parent Hulthemia is equally difficult to cultivate in our climate.

In view of these facts, there does not seem to be much hope of raising new races of garden Roses outside the genus Rosa of Linnaeus. In that genus, however, there is plenty of scope by introducing, as suggested, the chromosomes and characters of the E septet which are lacking in our present garden races.

Hurst bibliography