Euphytica 25 (1976) 361-365
Institute for Horticultural Plant Breeding (IVT). Wageningen, the Netherlands
Received 14 April 1975


In a three year's trial in the greenhouse, flower yields of seedling Hybrid Tea-roses both on their own roots and on a rootstock were recorded. Seedlings on their own roots yield similarly to plants on a rootstock. Highly significant correlations were found annually between the two categories. A selection method for potential yield on the basis of the first years yield of seedlings on their own roots, is described. Particularly in combination with selection for flower quality this method saves time, labour and space.


In N.W. Europe more than 500,000 rose seedlings are raised annually, a large percentage of which are tested as cut-roses. Traditionally, selection of seedlings for cut roses is done in two phases. At flowering in late spring, mass selection for mainly flower characteristics is practised. Every seedling with an acceptable flower is budded on a rootstock and, with dormant buds, planted early next year in a greenhouse. In that year the main part of the second phase of selection is carried out on relatively large numbers of small clones, with very few ultimately meeting the requirements.

As the above method of selection is rather cumbersome, it was considered whether direct budding could be avoided and the half year period following first flowering be used for a more rigid selection on plants on their own roots. At the end of that period, fewer, but better, seedlings could be grafted, while the plants were then in the same developmental stage as the budded plants in the first system.

Flower yield being one of the main demands in modern cut-roses, a trial was set up in which the first (half) years' flower yield of seedlings was compared with two years' yield of the same plants both on their own roots and on a rootstock.


The plant material consisted of 8 Hybrid Tea seedling populations, the progenitors of which were mated in 1970. After germination of the stratified seeds, most seedlings had flowered for the first time in May 1971.

Subsequently, from each of the 8 populations the first 25 large-flowered plants were chosen which flowered about simultaneously. These were planted in beds in a heated greenhouse. A two-row planting system was applied with about 8 plants per m2. The plants were grown and treated as cut-roses. At the end of the first growing season, in December 1971, each seedling was grafted on the rootstock 'Brögs Stachellose'. After a resting period at low temperatures of about 6 weeks and subsequent pruning of the plants on their own roots, the greenhouse was heated again and one individual of each grafted plant was planted. At that time each plant was in duplicate, a one-year-old on its own roots and a four-weeks-old vegetative descendant on a rootstock. Altogether 2 x 183 plants were involved.

In 1971, 1972 and 1973, from early February to mid-December, the number of cut flowers harvested from each plant was recorded. The trial ended in December 1973.


Table 1 presents the average annual flower yields of seedlings on their own roots in 1971, 1972 and 1973, and on a rootstock in 1972 and 1973. It also gives the coefficients of correlation between these parameters.

Cut flower yields of both categories increased with the ageing of plants. Owing to a six months' earlier start, plants on their own roots on average yielded slightly more flowers than those on a rootstock, but differences were not significant. Only in the six months season of 1971, plants on their own roots yielded significantly lower than in the next two years both on their own roots and on a rootstock.

The highly significant correlations between flower yields of plants on their own roots in 1971, and the annual yields recorded subsequently, indicate that the yields of seedlings on their own roots in the first half year, determined to a large extent the yields in later years, but the level may increase considerably.

Correlations between flower yields in 1972 and 1973 of plants on their own roots and those on a rootstock, also taking into account annual averages, indicate that plants on their own roots generally do not yield differently from plants on a rootstock.

Table 1. Average cut flower yield per plant (standard deviation between brackets) and coefficients of
correlation between the annual yields of 183 H.T.-rose seedlings, on their own roots and on a rootstock.

  Average flower
yield per plant
own roots rootstock
1971 1972 1973 1972 1973
Own roots 1971 6.3 (±3.3) - 0.60 0.60 0.52 0.45
Own roots 1972 20.7 (±11.5) - - 0.71 0.61 0.49
Own roots 1973 25.3 (±14.4) - - - 0.53 0.52
Rootstock 1972 18.5 (±10.8) - - - - 0.74
Rootstock 1973 21.3 (±12.3) - - - - -

1All correlations significant at P = 0.001.


Rose seedlings on their own roots appeared to develop and produce flowers similarly to plants on a rootstock, which is confirmed by others, comparing varieties of cutroses both on a rootstock and on their own roots. (ALLEN, 1970; ANON., 1973). This is advantageous to the selection procedure in seedling roses, because, most characteristics that determine the value of a cut-rose can be assessed in an early stage.

From variety trials with cut-roses, carried out at the Experimental Station for Floriculture at Aalsmeer, it can be derived that the second years' yield of a variety is characteristic of its production capacity, but that considerable differences are found between yields of small-, medium-, and large-flowered varieties.

The average yields in the second year, as calculated from data of VAN MARSBERGEN (1969, 1970, 1971, 1972, 1973). were for varieties in these three categories respectively 30.6 (σ = ±10.8, n = 16), 25.6 (σ = ±7.2, n = 25) and 17.5 (σ = ±5.7, n = 61) flowers per plant. Distributions of the categories over classes of yield are presented in Fig. 1. Particularly the large-flowered varieties, the most important category in cut-roses, yield relatively low. However, in 36, 7 and 2% of them, respectively ≥ 21, ≥ 26 and ≥ 31 flowers per plant were harvested, indicating that possibilities of improving the yield exist in this category. This is most urgent, because of the high production costs per unit area, which increases steadily with the costs of labour and energy. In view of the figures presented, clonal yields of more than 20 flowers per plant appear to be feasible. Selection for this yield level should preferably be done in an early stage of seedling evaluation, in order to prevent the propagation of plants not filling the requirements.

Considering the correlation (r = 0.45, n = 183) between the first years' yield of plants on their own roots and that in the second year on a rootstock, the question arose, what yield level selected seedlings should have in order to stand best chances of finding clonal yields of more than 20 flowers per plant.

Fig. 1. The distribution of small (S). medium (M) and large (L) flowered cut -rose varieties over classes of flower yield, 1 = 1-5, 2 = 6-10, 3 = 11-15, 4 = 16-20, 5 = 21-25, 6 = 26-30, 7 = 31-35, 8 = 36-40, 9 = 41-45, 10 = 46-50 flowers per plant in the second year on a rootstock. (After VAN MARSBERGEN). Arrows indicate averages.

Table 2 shows what percentage of seedlings, yielding in the first year on their own roots in classes 1-2, 3-4 ...., 17-18 flowers per plant, yielded in the second year on a rootstock in classes 1-5, 6-10, …., 61-65 flowers per plant. In the right margin of the table the percentage of plants is presented that yielded on their own roots ≥1, ≥3, …, ≥17 flowers per plant, and in the lower margin those that on a rootstock yielded ≥1, ≥6, …, ≥61 flowers per plant. For example, 69.2 % of the plants on their own roots yielded 5 or more flowers, and 81.1% of the plants on a rootstock yielded 11 or more flowers per plant. As to the aim of 20 flowers or more, it is seen that there is ample scope for selection, because 47.1%, 33.8% and 20.7% of the plants grafted on a rootstock produced ≥21, ≥26 and ≥31 flowers respectively.

Table 2. The relation between flower yields per plant in the first year of seedlings on their own roots, and that
of the same plants on a rootstock in the second year; figures are presented as a percentage of the total (110 = 183)

Yield in 1st year
on own roots
Yield in 2nd year on a rootstock
1-5 6-10 11-15 16-20 21-25 26-30 31-35 36-40 41-45 46-50 51-55 56-60 61-65  
1-2 1.6 2.2 1.6 0.6 0.6 0.6               100.0
3-4 1.6 5.5 3.8 6.0 0.6 4.9 0.6 0.6           92.8
5-6 2.2 3.8 3.3 4.9 4.4 2.7 2.7 1.1 1.1         69.2
7-8 0.6 0.6 2.7 4.9 2.7 3.3 3.3 2.7 - 0.6 0.6     43.0
9-10     2.2 2.2 2.2 1.6 3.3   0.6         21.0
11-12   0.4 0.6 0.6 1.1   0.6   0.6   0.6     8.9
13-14   0.4   0.6 1.1               0.6 4.4
15-16         0.6                 9.7
17-18                         1.1 1.1
  100.0 94.0 81.1 66.9 47.1 33.8 20.7 10.2 5.8 3.5 2.9 2.9 1.7  

In the following it will be considered how these percentages would have been, if successively seedlings that produced ≥3, ..., ≥11 flowers per plant, were retained for grafting. These figures which are derived from Table 2, are presented in Table 3. Obviously, in selecting seedlings of higher yield classes, the percentages of retained plants decrease, but also the percentages of valuable seedlings which are lost. The latter occurs sooner in plants with relatively low yields (≥ 21 flowers) than in those with high (≥ 31 flowers) yields. Retaining seedlings with too low or too high a yield is either ineffective or too rigorous.

Table 3. The effects of retaining seedlings above a certain yield class (≥1, …, ≥11 flowers per plant) determined during
the first year on their own roots, on the subsequent percentage of plants to be grafted and yielding ≥21, ≥26, or ≥31
flowers per plant in the second year on a rootstock. Percentages of valuable seedlings lost between brackets.

% of plants
% of retained seedlings
yielding in the second year
21 fl. 26 fl. 31 fl.
≥1 100 47 (0) 34 (0) 21 (0)
≥3 93 49 (1) 36 (1) 22 (0)
≥5 69 56 (9) 39 (7) 29 (1)
≥7 43 63 (20) 49 (13) 35 (6)
≥9 21 71 (32) 48 (24) 29 (15)
≥11 9 66 (41) 44 (30) 44 (17)

In this case a choice can be made between seedlings that yielded 5 or 7 flowers. A practical approach would be to retain the about 50% of seedlings that yielded equal to or slightly less than average (6.3 flowers), thus increasing the percentage of valuable plants considerably and keeping losses within reasonable limits.

In the above, selection for flower yield in otherwise unselected seedlings was considered. However, the method is much more advantageous when applied to those seedlings that traditionally would have been budded on a rootstock, viz. the seedlings that passed mass selection at first flowering. In that ease a combined selection for flower quantity and quality can be done, resulting in the choice of about the same valuable individuals after one year, as in clones after two years.

It is realized that the selection method described has consequences for management but particularly with respect to time, space and labour, it should be given consideration by the breeder of cut roses.