Euphytica 27: 205-210 (1978)
Institute for Horticultural Plant Breeding (IVT), Wageningen, the Netherlands
Received 10 August 1977


To transfer the yellow flower colour of R. foetida to Hybrid Tea roses, F1’s and backcrosses with Hybrid Tea's were produced. In B1 populations, yellow, recurrent flowering seedlings occurred. Recurrent flowering was controlled by one recessive gene.


Yellow coloured flowers were unknown in European roses before the introduction of 'Parks Yellow Tea-scented China' in 1824. However its yellow was rather pale and breeders were unsuccessful in intensifying it. It was not until 1898 that a real breakthrough towards deep yellow was made, when in France Pernet-Ducher of Lyon, after a few thousand unsuccessful pollinations, at last obtained two F1 seedlings from the cross 'Antoine Ducher' x R. foetida 'Persian Yellow'. From a backcross with an unknown Hybrid Tea cultivar as the female parent, he selected 'Soleil d'Or' (HURST, 1941). This tetraploid cultivar was the first of the Pernettiana class of roses later to be fully merged with the Hybrid Tea's. PARIS & MANEY (1944), in a comprehensive study, demonstrated that all the yellow cultivars then known, had 'Soleil d'Or' as a common ancestor. Our own literature study showed that this situation did not change in the following years. This means that the genetical basis of yellow in present-day roses is extremely narrow, and that it would be worthwhile to re-use R. foetida.

Although some authors (ANDERSON, 1969; WRIGHT, 1965) emphasized the value of R. foetida for modern breeding, to our knowledge no successful crosses have been made since Pernet-Ducher. Our present paper reports on such crosses made between 1971 and 1976.


The R. foetida Hermann material consisted of the cultivars Austrian Briar and Persian Yellow, obtained as pollen and grafts from the Agricultural University Arboretum at Wageningen in 1971. The white flowered Hybrid Tea (HT) cultivars used were White Weekend, White Masterpiece, Sincera and Pascali. All R. foetida (ANONYMOUS, 1969) and HT cultivars were tetraploids (2n =4x = 28). The R. foetida cvs flower only once a year and form large shrubs, the HT cultivars flower recurrently and have a relatively small (= dwarfing) habit. From 1972 the progenitors were planted in a coldhouse with supplementary heating in spring.

The main line of the breeding scheme was as follows (see also Table 1). Summer 1971 'White Weekend' was pollinated with 'Persian Yellow' and 'Austrian Briar'. Plants of the resulting F1's were selfed in 1973 and 1974, and in 1974 also pollinated with a mixture of F1 pollen. When they produced no seeds, in 1974 the remainder of the pollen mixture of F1 plants was used on 'White Weekend', 'White Masterpiece' and 'Pascali', resulting in a first back-cross (B1) to HT. Some back-crosses with the best F1 plants were repeated in 1975 and 1976. In B1 populations selection was made for yellow, recurrent flowering seedlings. To check the genetical background of flower colour of the original parent cultivars used, plants of these cultivars were selfed in 1972.

The number of pollinations made each year was limited by the numbers of flowers on the female parents. All pollinations were done with a No. 5 aquarel brush. Selfings were done just after anthesis, cross-pollinations were made with pollen stored in petri dishes in a dessicator at 1°C (VISSER et al., 1977) on flowers emasculated 24 hours previously.

Seedlings were raised from seeds stratified for 4 months at 0°C. They were grown in square 6x6 cm flower pots in a heated greenhouse (20°C), till flowering or nonflowering could be assessed. Non-flowering F1 seedlings were transplanted to a coldhouse till flowering occurred.

In this study plants that flower through the year are called: recurrent, those that flower only once: non-recurrent. Seedling populations are designated by five figures, of which the first two indicate the year, and the last three the sequence number of the cross (e.g. 71.503).


Results of crosses, as to numbers of seeds, seedlings and percentages of flowering, are presented in Table 1. It can be seen that the initial F1's (71.501 and 71.503) contained a total of 37 seedlings of which none flowered in the year of germination. All seedlings from selfed 'White Weekend' and 'White Masterpiece' (72.130 and 72.173 resp.) were recurrent and bore white flowers. 'Sincera', 'Persian Yellow' and 'Austrian Briar' either after selfing or in cross-pollinations, failed to set.

The F1 seedlings were relatively uniform as to vigour, showing a decidedly R. foetida habit. Thus leaves were composed of 7-11 leaflets (3-5 in HT) which had, like R. foetida, the typical smell of sour apples. Shoots were light red-brown, with very many fine straight thorns. When they flowered in the second or third year after germination, plants were well over 2 m high. Most flowers were intensively yellow and differed per plant in number, shape, size and petal number. Only plants with flowers most resembling HT's were selected for further breeding.

As in the R. foetida cultivars, cross- and self-pollinations of F1's failed to set.

Table 1. Review of successful crosses and self-pollinations of Hybrid Tea and R. foetida cvs from 1971 -1976.

Cross Year and
of flowers
of seeds
of seedlings
% seedlings
in1st year
% seedlings
in 2nd year
White Weekend x R. foetida ‘Persian Yellow' 71.501 12 3 1 0 100
White Weekend x R. foetida 'Austrian Briar' 71.503 12 70 36 0 100
White Weekend x self 72.130 6 181 80 100 0
White Masterpiece x self 72.173 6 160 61 100 0
White Weekend x 71.503 (poll. mix.) 74.570 49 2513 856 15 0
White Masterpiece x 71.503 (poll. mix.) 74.571 38 981 231 18 0
Pascali x 71.503 (poll. mix.) 74.572 16 222 54 13 0
White Masterpiece x 71.503 75.505 10 42 10 20 0
White Masterpiece x 71.503 75.507 6 146 44 11 0
White Weekend x 71.503 75.515 39 1254 214 13 0
White Weekend x 71.501 76.504 81 2085 682 17 0
White Masterpiece x 71.501 76.513 91 671 107 14 0


Fig. 1. B1 seedlings about 50 days old. Right: recurrent flowering seedling. Left: non-recurrent it flowering seedling.

However, pollen of F1's induced fruitset in HT's (1974, 1975, 1976). The better female parents in backcrosses with respect to fruitset, numbers of seeds, and flower shape in their offspring, were 'White Weekend' and 'White Masterpiece'. 'Pascali' was not used any more because most of its descendants had pinkish flowers.

B1-populations segregated into flowering and non-flowering seedlings. All flowering seedlings appeared to be recurrent. Their frequency depended on the population and varied from about 1l-20. Only B1's with yellow flowers were selected for future breeding. The non-recurrent B1's, which all closely resembled the F1's, were discarded.


In F1's of HT x R. foetida non-recurrent flowering seedlings only occurred, implying that non-recurrent flowering is dominant over recurrent flowering. This agrees with the suggestions made by VON RATHLEF (1928), LAMMERTS (1945), Swim (1948), FISHER & MOREY (1963) and MOORE (1968). SEMENIUK (1971a; 1971b) demonstrated that in both diploid and tetraploid descendants of 'Goldilocks' x R. wichuraiana, recurrent flowering was controlled by one recessive gene: r. If this applies to our HT varieties as well, they would be nulliplexes (r4). If R. foetida, which flowers only once a year, should be a quadruplex (R4), then the F1 is R2r2. Consequently in B1 populations the ratio recurrent : non-recurrent flowering will be 1:5, if random chromosome assortment is assumed (ALLARD, 1960). The actual numbers of recurrent and non-recurrent flowering seedlings occurring in each of the larger B1-populations agreed very well with this ratio (see Table 2).

It is therefore concluded that recurrent flowering in F1 and B1 populations with HT's and R. foetida cvs is controlled by one single recessive gene. According to SEMENIUK (1971b) recurrent flowering in 2x, 3x and 4x roses is controlled by one gene: r. We assume that this gene is also operative in the HT-cultivars used by us. This could mean that the dominant allele R, found by Semeniuk in the climbing R. wichuraiana, is identical with that occurring in R. foetida, suggesting a relationship between these species.

Recurrent flowering seedlings all follow the same developmental pattern, i.e. when 5-6 leaves are formed, the apical flower bud appears about 30 days after germination (DE VRIES, 1976).

Table 2. Comparison of actual and expected (1:5) numbers of recurrent and non-recurrent
flowering seedlings in B1's from crosses between HT's and F1 (HT x R. foetida)'s.

    Actual numbers
of seedlings
Expected numbers
of seedlings
Year and
B1-cross recurrent
74.570 Wh. Weekend x 71.503 131 725 143 713 0.29*
75.515 Wh. Weekend x 71.503 28 186 36 178 0.18*
76.504 Wh. Weekend x 71.501 119 563 114 568 0.65*
74.571 Wh. Masterpiece x 71.503 41 190 38 193 0.65*
75.507 Wh. Masterpiece x 71.503 5 39 7 37 0.40*
76.513 Wh. Masterpiece x 71.501 15 92 18 89 0.42*
74.572 Pascali x 71.503 7 47 9 45 0.49*
* Differences significant for P<0.05.          

In populations, non-recurrent seedlings can be easily distinguished because they do not form a flower bud in the first year. In a coldhouse they do so in April next year. It is yet unknown what factors determine flower initiation in this case.

It is concluded that the transmission of yellow flower colour to recurrent flowering Hybrid Tea roses can be realized in two generations of breeding. The same method may be applied to transfer genes not yet available in modern varieties.


The authors wish to thank the late Mr G. Verbeek, rose breeder at Aalsmeer, for kindly placing 'White Weekend' at our disposal for use in the initial crosses.