Chromosome Botany (1956, 1963, 1973)
C. D. Darlington

6. Rosa

The 178 species of the genus included in the Chromosome Atlas are distributed throughout the North Temperate Zone. They provide one of the classical examples of a polyploid series as shown by the work of Täckholm in 1922. The basic number of seven is common to them all. Departures from strict balance are rare. They reach their limit of multiplication in nature with an octoploid, the circumpolar R. acicularis.

The genus contains one genetically aberrant group: the dog roses which were recognized as the Caninae section in 1863. They are native throughout Europe and Western Asia. They have been allowed eighteen species but the allowance is quite arbitrary and in fact the whole group in its whole area of distribution is in a state of flux. It includes innumerable forms which inter-breed in nature in some areas but in others are isolated and therefore largely self-fertilised. The only substantial basis of division is by chromosome number into three classes, 4x, 5x and 6x.

The reason for these conditions is chiefly the versatile system of reproduction. The basic system is subsexual but of a unique kind: 7 pairs of chromosomes are formed at meiosis together with a residue of 14, 21 or 28 univalents. The univalents are lost in pollen formation, but carried unreduced by the eggs so that the self-fertilized plant breeds approximately true and, on crossing, its genes are recombined only in a small part of its chromosomes, the seven pairs. Thus, the chromosome complement may be represented as AA (7II) BCD (21I) in a pentaploid Canina rose. The A group is sexual and bi-parental. The BCD group is exclusively maternal in inheritance and vegetative or clonal in character. And the two groups, the two parts of heredity, are strictly separated in evolution.

Subsidiary to this subsexual system is a purely vegetative apomixis which probably competes with varying success in different individuals and species with the results of selfing and crossing. These peculiarities of the Caninae group have not yet been introduced into the main classes of garden roses which all seem to be normally sexual.

Garden Roses

Since Minoan times in Europe, and for nearly as long in China, roses have probably been grown in cultivation. Double-flowered forms have been found in the wild in over twenty species in Europe and China and recently in America. These especially have been selected and bred for decoration and, in the case of the damask roses, for perfume as well. It was only after the year 1802 that systematic crossing began to yield the great improvements from which modern roses have been derived. These have all come from seven Old World species, or cultivated groups to which, by an easy-going tradition, we give the names of species. The history of the crossing and improvement of the resulting garden roses is marvellously preserved for us, notably by botanical illustrations (which confirm our names), but also very often by the preservation of the critical plants in living collections. These have revealed that hybridization and selection have been carried out on a basis (unknown at the time) of several degrees of polyploidy. From the determination of the chromosome numbers in the chief groups it has therefore been possible to understand, to correct, and to explain the whole history.

The seven important species fall into two groups which are contrasted in their polyploidy, their geography and partly also in their habit. There are the diploids from the east which are all tall bushes or climbers. And there are the tetraploids from the west which are all smaller bush roses.

Three of these species, all eastern diploids, R. gigantea, R. multiflora and R. wichuraiana, present no great difficulty. They occur in the wild as white-flowered single forms. Improved, that is, double and coloured forms of multiflora had appeared in China while similar forms of gigantea had contributed to chinensis. These species, however, were only introduced into the history of crossing after 1800. The other species need to be separately considered:

(1) R. chinensis (2x). This is the old rose of cultivation in China no longer connected with any single wild species. It is, unlike the three wild species, a bush rose with several colours of flowers. In cultivation, no doubt after hybridization, it has been selected for double flowers of varied colours and strong scent (hence 'R. odorata') and for a perpetual flowering habit (hence 'R. semperflorens'). Its earliest record in Europe according to Hurst is in a painting by the Florentine Bronzino dated 1529. Being less hardy than the European roses, it did not appear in England until it came direct from China in 1789.

(2) R. moschata (2x). The Musk Rose grows wild in the Himalayas but already in prehistoric times its double forms were cultivated and had spread east to China. It had also spread west to Greece by the time of Theophrastus. It reached England in the reign of Henry VIII.

(3) R. gallica (4x). The 'centifolia' or hundred-petalled rose of classical times and the French Rose of the Middle Ages grows wild in single forms in southern Europe and the Levant. It lends its name to the deep red roses of Lancaster and also of Tuscany. Probably by hybridisation in the Mediterranean countries with R. moschata (which would double its chromosome number to give fertile tetraploid hybrids), this group gave the Damask Roses already in ancient times. They penetrated into England along with the Musk Roses. But the two, which had crossed in the east much earlier, in the west were now quite independent: perhaps conditions of pollination had changed. From the Damasks in turn the Cabbage Roses, the modern R. centifolia, were developed to the number of 2,000 varieties by Dutch breeders in the seventeenth century (1580-1710).

Modern roses with their great variety of colour, shape, habit and hardiness began to appear when the barrier between the first of these species, chinensis, and the other two, moschata and gallica was broken down by crossing, partly deliberate crossing. This happened in a series of steps as follows (cf. Table 17):

(i) R. chinensis (2x) x R. moschata (2x). The NOISETTE roses of tall bush habit were produced as the F2 from this cross at Charleston, U.S.A., in 1802.

(ii) R. chinensis (2x) x R. damascena (4x). The BOURBON roses of the expected bush habit were produced as a spontaneous 3x F1 whose occasional 2x gametes gave 4x plants in a back-cross to the R. gallica group. They were raised on the Ile de Bourbon in 1817.

(iii) From the same types of crosses but using various members of the gallica group the more heterogeneous Hybrid China roses were raised in France and England in the years following 1820.

The second half of the nineteenth century saw the development of two further distinct groups or families of roses, two further steps in mixing. First, there were the diploid Tea roses, dwarf segregates from Noisette and R. gigantea crosses. They had flowers of exquisite form, borne all the year round in favourable climates, but were not very hardy. Secondly, there were the tetraploid HYBRID PERPETUALS from Bourbons and Hybrid Chinas crossed back with the European gallica varieties. The Hybrid Perpetuals were not as good as their name. They naturally had only a weak second flowering season but they were hardier than the Hybrid Chinas. These two groups included the bulk of the roses grown in the period 1850 to 1890. From 1890 onwards, the five old groups began to be superseded by a sixth group, the HYBRID TEAS, produced from crosses between these last two and combining the merits of their parents. It was in this group, we may note, that the chromosomes of the three great parent stocks moschata, gallica and chinensis were effectively brought together for the first time, brought together, that is to say, in fertile hybrids. The original Hybrid Teas were again, of course, triploid, but again tetraploids arose by non-reduction in the course of crossing the less sterile triploids back with the Hybrid Perpetuals.

Further variation was brought into the Hybrid Tea group at the beginning of this century from the seventh species, the yellow and copper tetraploid R. foetida, in a triple cross. This supposedly Persian species is not known in nature and is itself doubtless, like chinensis, a cultivated group of ancient hybrid origin. Its bicolor mutation is responsible for the brilliant yellow, orange, and coppery tones first found in the PERNETIANA group.

Another line of advance began at the same time as the Hybrid Teas. It was purely eastern, purely diploid and therefore less important in its immediate results. It came by crossing the other diploids with the wild multiflora of Japan and gave the POLYANTHA roses. These arising from a triple cross in the fourth and later generations were selected in two directions, dwarf and climbing. Unlike the Hybrid Teas, their flowers were very small, and borne in large clusters. The Dwarf Polyanthas, moreover, had a very long flowering season. Many varieties arose later, some of them as bud mutations. The mutation to the formation of pelargonidin came first in the late 1920's to give the diploid Paul Krampel (Scott-Moncrieff, 1936). More recently it gave among its derivatives the tetraploid Sondermeldung or Independence (Dayton, unpub.). Pelargonidin is a pigment unknown in the ancestral species and indeed new for the whole genus.

The most recent group of garden roses, the HYBRID POLYANTHAS, came from crossing the Dwarf Polyanthas with Hybrid Teas. The first varieties were again triploids and were sterile with a few exceptions. From these the later tetraploids have arisen, principally through crossing back to the Hybrid Teas. But there are varieties similar to the true Hybrid Polyanthas in their clusters of medium-sized flowers borne over a very long period, and these have moschata or wichuraiana, in addition to R. multiflora, in their ancestry.

Bud Sports. Among these crosses, both 2x and 4x climbing or rambler varieties have appeared from time to time. They have all been derived from crosses with the four tall or climbing species. Successively the Noisette, Hybrid Tea, and Climbing Polyantha groups have been most important: and last of all Hybrid Wichuraiana have predominated by reason of their disease resistance and longer flowering. One special feature of the climbers is their frequent appearance in the Hybrid Tea bush roses by bud mutation. These have frequently reverted to the bush type possibly owing to their having arisen as chimaeras.

Another example of bud mutation giving rise to a chimaera is the scarlet rose Paul Krampel which when eight years old gave pink and crimson bud sports on the same plant. Probably the new form had arisen as a chimaera of old and new types and was giving the older pigment types by reassortment of tissues.

The largest scale of bud mutation ever invoked in plants perhaps is that by which double Moss Roses, with resinous hairs on the fruit, according to Hurst, came from the Cabbage Roses during the eighteenth century. Over sixty clonal varieties are known. All of these are sexually sterile and several of them have been known to revert in cultivation to the Cabbage Rose type. Again, therefore, these mutations may have existed as unstable chimaeras. The precise internal and external conditions of bud-mutation evidently deserve special study in Rosa.

Stocks. In ancient times the two main groups of roses, gallica in Europe, odorata in China, could be readily propagated by suckers or cuttings. Grafting experiments were no doubt attempted, but it was only when new gallica forms arose in the Middle Ages which were difficult to multiply directly, that gardeners would resort to using the borrowed roots of older garden forms without such a defect.

Suitable stocks were no doubt selected in this way by nurseries during the eighteenth century. At this time, too, pentaploid Caninae roses began to be used, having the advantage that they could be gathered as wild seedlings. In the nineteenth century three new groups of roses began to be grown and even bred for stocks: (i) the recently introduced diploid species, multiflora and rugosa; (ii) selections from their hybrid progeny; and (iii) established garden varieties such as manetti, a supposed Noisette raised in 1837. The last included not only diploids but also triploids, and tetraploids: differences of chromosome number proved to be no barrier to grafting.

Conclusion. The garden roses have been derived not from the hexaploids and octoploids but entirely from the lower-numbered species. The largest section of the genus, the Cinnamomeae, which contains the greatest development of polyploidy, has taken no part in the great programme of Table 21. What garden forms it has produced came by mutation and hybridization from its diploid species, rugosa and cinnamomea.

Moreover, such increases of chromosome number as have occurred have been solely to raise the progeny of these crosses to a uniform tetraploid level. Higher polyploids must certainly have arisen in the course of these innumerable crossings, as they have in Narcissus. But they have evidently been rejected by breeders. It is at the tetraploid level that interspecific hybridisation has been able to release the great diversity that we now see. It has worked largely by recombination but partly also by the selection of instability. And its future possibilities for the genus are still beyond estimation.