Euphytica 23 (1974) 665-669
1Contribution No. 73/2.
Canada Agriculture, OTTAWA, Ontario, Canada Received 18 January 1974


Reciprocal crosses were carried out for 6 combinations of 4 R. rugosa with 3 R. chinensis cultivars. The parental combinations were evaluated for successful pollinations, seed germination and flowering and seed setting seedlings in populations of 3 year old plants. Seeds and viable seedlings were obtained from each of the 12 seedling populations. Only 3 populations produced female fertile offssprings. The average of seed setting seedlings was 1.5%. The percentage of flowering seedlings was significantly reduced in the offspring from certain parents. The success of cross pollination, seed germination and the ability to initiate flowers were influenced by cytoplasmic factors.

The differences in fertility between the present hybrids and the only other rugosa x chinensis hybrid described, R. calocarpa (ANDRÉ) WILLMOTT, are discussed.


The crosses between Rosa rugosa and R. chinensis cultivars were carried out to obtain hybrids from which new types of everblooming and winterhardy roses could be developed. Contemporary garden roses are everblooming but are not winterhardy in most parts of Canada. They can be grown without winter protection only along the coast of British Columbia and in the most southern region of the Niagara Peninsula. Native rose species, which survive the winters without protection in the colder areas of Canada, are not everblooming.

The everblooming habit of contemporary roses derived from roses which were introduced into Europe from China and India during the second half of the 18th Century. These came from cultivated plants. Also, nearly all herbarium specimens including the rose described by JACQUIN in 1768 as R. chinensis derived from cultivated plants (WILLMOTT, 1910-1941).

R. rugosa THUNB., a native of Japan, is as hardy as the Canadian species. It is well adapted and is frequently found as an escape. Its outstanding feature among hardy species is the ability to flower repeatedly during one season. Many hybrid of R. rugosa are known but, to my knowledge, only one hybrid of R. rugosa with R. chinensis was described previously. The cross was made by Bruant of Poitiers before 1891 and one of the resulting seedlings was described by ANDRÉ as R. calocarpa (WILLMOTT, 1910-1914).

Through interspecific hybridization revolutionary changes are often obtained in ornamental plants but the breeding of interspecific hybrids is often limited by reproductive barriers. These cover a wide range of hybrid sterility and inviability. Hybrid sterility is most frequently caused by the abortion of pollen or ovules. Hybrid weakness or inviability is another form of reproductive failure when the plant is unable to complete certain stages of development essential for its own survival or for the survival of its kind. Normal appearing seeds might fail to germinate or the seedlings might die soon after germination. Some seedlings grow well but are unable to initiate flowers. Even if one of the F1 seedlings is able to set seeds, the seeds might not germinate or the second generation of seedlings might be sterile.


All cultivars used are diploid with 14 somatic chromosomes. The pollen parents of most of the rugosa and chinensis cultivars are not known, but it can be assumed that the rugosa cultivars originated from crosses between Hybrid Tea and Floribunda and that the chinensis cultivars originated from crosses between roses introduced from China and India with roses of 18th Century Europe. Since the chinensis cultivars are ancestors of the Hybrid Tea and Floribunda, it could be assumed that the rugosa and the chinensis cultivars were more closely related than their parent species. The assumption that a closer relationship of the parent would facilitate the recovery of fertile hybrids was the reason why cultivars were used instead of the parent species

*CybeRose note: I would be interested to know the source of this date. Others I have seen are references either to the Monthly Damask, or to Rosa indica L, which was probably intended for R. banksiae.

The cultivars of the rugosa group were: Frau Dagmar Hartopp, Martin Frobisher, Schneezwerg and Will Alderman. The cultivars from the chinensis group were: Duchesse de Brabant, Le Vésuve and Old Blush. Frau Dagmar Hartopp is also list as Fru Dagmar Hastrup, a low growing shrub, whose parents are not known. Martin Frobisher derived from open pollination of Schneezwerg (SVEJDA, 1969). Schneezwerg is a hybrid of rugosa parentage. Different authors agree that it resulted from a wide cross but they are uncertain of which species was the pollen parent. According to the International Check-List of Roses, it is presumed to be a Polyantha hybrid (MCFARLAND, 1969). THOMAS (1963) suggests R. bracteata and WYLIE (1954) suggests R. multiflora. Will Alderman derived from a cross of (R. rugosa x acicularis) x a Hybrid Tea (MCFARLAND, 1969). Duchesse de Brabant is one the early hybrids of R. chinensis which has been cultivated since 1857 (MCFARLAND, 1969). Le Vésuve has been cultivated since 1825. STEEN (1959) wrote that this rose is, according to some books, identical to R. cruenta of Redouté but thought it resembles 'Crimson China'. The Crimson China rose is botanically R. chinensis semperflorens (CURTIS) KOEHNE. Old Blush or Common Monthly was an early introduction from China. It came to Sweden in 1752* (MCFARLAND, 1969).

Reciprocal crosses were carried out for all parental combinations. Some cultivars were used repeatedly but complete diallel series were not obtained.

The reproductive capacity of the hybrids was evaluated on the proportions, (a) successful pollination, (b) germination resulting in viable seedlings, (c) flowering seedlings in populations of 3 year old plants, (d) female fertility of 3 years old plants.

With one exception, each seedling population comprised a considerable number of individuals, varying between 54 and 901. The one exception was the combination of Martin Frobisher with Le Vésuve from which only 8 seedlings were obtained.


Crossability. The proportions of successful pollination between parental combinations varied but seeds were obtained from each. Thus, incompatibility between parents was not found. Significant differences were obtained between the reciprocal crosses of Will Alderman with Le Vésuve and of Schneezwerg with Old Blush (Table 1). The difference between the reciprocal crosses of Will Alderman with Le Vésuve might have been a cytoplasmic effect. The difference between the reciprocal crosses of Schneezwerg with Old Blush might have been due to the fewer ovaries in the receptacle of old Blush since the cross was less successful when Old Blush was the pistillate parent. In the first case, the cross was more successful when the cultivar with the lower number of ovaries, Le Vésuve, was the pistillate parent.

Table 1. Reciprocal differences in crossability and in characters of the F1 generation in combinations of R. rugosa (A) with R. chinensis (B) cultivars.

Parental combinations Pistillate parent Difference
A B A B  
% successful pollinations
Will Alderman Le Vésuve 3.7 46.8 43.1*
Martin Frobisher Le Vésuve 47 14.6 9.9
Schneezwerg Le Vésuve 36.6 11.6 25.0
Schneezwerg Old Blush 454 5.6 39.8*
Schneezwerg Duchesse de Brabant 44.6 39.8 4.8
Frau Dagmar Hartopp Old Blush 9.6 10.1 0.5
% germination
Will Alderman Le Vésuve 2.0 23.2 21.2*
Martin Frobisher Le Vésuve 13.0 6.6 6.4
Schneezwerg Le Vésuve 6.6 5.2 1.4
Schneezwerg Old Blush 6.2 3.5 2.7
Schneezwerg Duchesse de Brabant 12.2 8.0 4.2
Frau Dagmar Hartopp Old Blush 43.3 7.5 35.8
% flowering seedlings
Will Alderman Le Vésuve 25.0 30.0 5.0
Martin Frobisher Le Vésuve 100.0 50.0 50.0
Schneezwerg Le Vésuve 46.0 0.0 46.0**
Schneezwerg Old Blush 81.2 28.6 52.6*
Schneezwerg Duchesse de Brabant 38.6 56.6 18.0
Frau Dagmar Hartopp Old Blush 80.6 80.0 0.6
% seed setting seedlings
Will Alderman Le Vésuve 0.0 0.0 0.0
Martin Frobisher Le Vésuve 0.0 0.0 0.0
Schneezwerg Le Vésuve 0.0 - -
Schneezwerg Old Blush 1.5 0.0 1.5
Schneezwerg Duchesse de Brabant 1.6 4.3 2.7
Frau Dagmar Hartopp Old Blush 0.0 0.0 0.0
*Difference significant at P ≤ 0.05.
**Difference significant at P ≤ 0.01.

Seed germination. The means of germination varied between 4 and 45%. The highest mean was obtained from the cross of Frau Dagmar Hartopp x Old Blush and the lowest for the cross of Old Blush x Schneezwerg. A significant difference between reciprocals was obtained for the combination Will Alderman with Le Vésuve. Seed germination was increased when Le Vésuve was the pistillate parent. It is interesting to note that, for this cross, both successful pollinations and seed germination were increased when the CHINENSIS cultivar Le Vésuve was the pistillate parent (Table 1).

Development of flowers. The percentages of flowering seedlings in 3 year old plants varied between 0 and 100. Flower bud development was influenced by cytoplasmic factors as shown by the differences between reciprocal crosses in the combinations Martin Frobisher with Le Vésuve, Schneezwerg with Le Vésuve and Schneezwerg with Old Blush (Table 1). Significantly fewer seedlings flowered when the CHINENSIS cultivar was the pistillate parent in the 3 combinations.

Seed setting. Only 14 of the 2047 hybrid seedlings produced seeds after free pollination. Seed setting seedlings were obtained in 3 populations from the cross of Schneezwerg with Old Blush when Schneezwerg was the pistillate parent and from the crosses of Schneezwerg with Duchesse de Brabant regardless of which cultivar was the pistiliate parent (Table 1).


The high level of hybrid sterility in the offspring from our crosses between rugosa and chinensis cultivars showed a reproductive barrier not suggested by the hybrid obtained by Bruant of Poitiers which was, as the name calocarpa implies, notable for its beautiful fruit. Surprisingly, it was found among a few seedlings from a cross of R. rugosa with a common China Rose (WILLMOTT, 1910).

It is known, but not always accepted as fact, that the fertility level of the hybrid depends on the choise [sic] of parents. The fertility level of the offspring from a certain parental combination cannot be predicted. It must be found by trial. Irregularities at meiosis might lead to the abortion of pollen and ovules but this is only one of several causes for hybrid sterility. It is therefore not surprising that a greater number of fertile hybrids was not obtained from cultivars.

The difference of the fertility levels of the hybrids from R. rugosa obtained by LEWIS (1961) and by ERLANSON (1931), shows that it is difficult to duplicate the results from interspecific hybridization experiments. The hybrid of Lewis showed 16.7% defective pollen grains but the hybrid of ERLANSON was 100.0% defective. Since the hybrid of ERLANSON was not hardy at Ann Arbor, Michigan, the absence of pollen fertility could have been also an effect of the environment but this did not rule out genetic differences (LEWIS, 1961). ERLANSON (1931) showed also that pollen sterility varies between individual plants of the same species and she concluded that genetic factors were at least partially responsible for it.

Regrettably, R. calocarpa is no longer obtainable, its noted ability to produce abundance of hips is astonishing since very few of my rugosa x chinensis hybrids were able to set even a few hips. The hybrid character of R. calocarpa (WILLMOTT, 1910) was shown morphologically by features from both parents. Roses hybridize freely in their natural habitat (ERLANSON, 1929) and in controlled cross breeding. Even complex hybrids may produce an abundance of hips. The rugosa hybrids Schneezwerg and Will Alderman are examples for the high fertility of certain complex hybrids.

The hybrid character of the seedlings from rugosa and chinensis cultivars was shown most clearly by the level of winter survival. The rugosa parents and the offspring from open pollination survived the winters at Ottawa without injury or with tip killing only. The chinensis parents and the offspring from cross pollination within the group did not survive winters outdoors. The hybrids from rugosa and chinensis survived winters outdoors but in most cases the wood was killed back to the snowline. The morphological characters varied but all hybrids had features from both parents. The styles of the chinensis are exerted, but the styles of the rugosa are not. In the hybrids, exertion of styles prevails. The number of leaflets of the rugosa parents varies from 5 to 7 and of the chinensis parents from 3 to 6. The number of leaflets of the hybrids varied from 3 to 9 but individuals with 5 to 7 leaflets prevailed. It should be added that winter survival was noted for each hybrid seedling but morphological features were noted only for selected seedlings. Since exertion of styles and number of leaflets were not selection criteria it can be assumed that the selected seedlings constituted a random sample for these characters.

The influence of the cytoplasm, as indicated by the significant differences between the reciprocal crosses of some hybrid populations, on the initiation of flowers, on the success of cross pollination and on seed germination is interesting because it shows that the success of crosses and successive seedling development is not entirely dependent on the genetic constitution but on the harmonious interaction between genes and cytoplasm.

It is possible to obtain new types of winterhardy and everblooming roses from the rugosa x chinensis hybrids, but this is an elaborate and time consuming process. During the progress of this work I found shorter ways to obtain the desired results.


It is a pleasure to thank Messrs. N. Henry and G. Rofner for carrying out the large number of cross-pollinations for this study. I would also like to thank Drs B. Boivin and S. Fejer for helpful criticism of the manuscript.