Euphytica 26: 761-764 (1977)

Institute for Horticultural Plant Breeding (IVT, Wageningen, the Netherlands
Received 18 March 1977


In large flowered rose seedlings of the Hybrid Tea type, very significant positive correlations were found between:


In cut roses the length of the cut stems plays an important role. The wholesale trader pays better prices for long stemmed roses than for short ones. This is true both within and between varieties. Because long stemmed (= large flowered) varieties yield fewer flowers per plant than short stemmed (= small flowered) varieties (DE VRIES, 1976b), the former category does not necessarily give the grower better returns per unit area. Nevertheless there is a tendency to grow large flowered varieties rather than small flowered ones.

This demand for large flowered varieties should be a stimulus for the breeder to improve their yielding capacity.

Previously DE VRIES (1976a) showed that Hybrid Tea-rose seedlings with short juvenile periods produce more flowers than those with along juvenile period. Because the flower production of own rooted seedlings is closely related to their production when grafted on a rootstock (DE VRIES, 1976b) the juvenile period is likely to be an important criterion in the selection of better yielding varieties.

Our present study deals with the early selection for stem length in rose seedlings of the Hybrid Tea type, with respect to the juvenile period.


The plant material consisted of 3 different batches of hybrid Tea (HT)-rose seedlings. The first batch, hereafter referred to as 1973-1, consisted of 6 seedling populations the parents of which were mated in 1973. After stratification and subsequent germination (see also 'Discussion') of the seeds in March 1974, on the first 31 large flowered seedlings of each population the following data were recorded: date of germination, date of flower bud appearance, date of first flowering and plant length at first flowering. In January 1975 each plant was grafted onto the rootstock Brögs Stachellose and 4 weeks later planted in a heated greenhouse. Altogether 186 plants were involved. From September to November 1975, when the plants were fully developed, the lengths of 5 shoots of each plant were measured from their base to the apex when the flowers were about ripe to cut.

The second batch of seedlings, indicated by 1973-2 consisted of the 36 most promising large flowered seedlings of the 1974 selection season. Their length was measured in 1974 at first flowering. In January 1975 each seedling was grafted onto Brögs Stachellose and planted at six plants per clone in the same greenhouse as the first batch. Shoot lengths of these grafted plants were determined similarly to those of the first batch.

The third batch of seedlings, designated as 1974-1 consisted of the 27 most promising large flowered plants of several 1974 populations. The seedlings were grafted on Brögs Stachellose in January 1976, and observations on stem-length were the same as for the first and second batches.

The juvenile period (JP) of seedlings is defined as the number of days from germination to flower bud appearance (DE VRIES, 1976a).


Fig. 1 presents, for the three batches of seedlings, the relation between the length of the seedling at first flowering and the shoot length of the same plants when grown on a rootstock, expressed as a percentage of average length per batch. Correlations between the two parameters were for 1973-1: r = 0.66 (n = 186); 1973-2: r = 0.70 (n = 36); 1974-1 : r 0.68 (n = 27). All correlations were significant for P = 0.001.

Fig. 2 presents the relation between the juvenile periods of the 1973-1 batch of seedlings and their shoot length as grafted plants. Because the two parameters are positively correlated (r = 0.88, n = 35), longer JP's of the seedlings are associated with longer shoots on the grafted plants.

Fig. 1. The relation between seedling length at first-flowering and relative shoot length of the same plants grafted onto a rootstock, for 3 batches of HT-rose seedlings (mean shoot length 1973-1 = 55 cm; 1973-2 = 73 cm; 1974-1 = 60 cm). Fig. 2. The relation between the juvenile periods of HT-rose seedlings and relative shoot lengths when grafted on a rootstock.


Notwithstanding pre-selection for flower performance in two batches of seedlings (1973-2, 1974-1), highly significant correlations were found between the parameters concerned.

From the positive correlations between shoot length of the young seedling and that of the grafted plant (Fig. I), a relatively simple method of selection for stem length may be derived.

According to the Table of the bivariate normal distribution-function (FERGUSON, 1965), discarding seedlings with a below average shoot length at first flowering will give grafted plants of which about 80% have an above average shoot length. This selection may be made even more efficacious by taking the JP of the seedlings into account, because long JP's generally go with long stems.

However, apart from long stems, high production in cut roses is demanded. In young seedlings few characters appear to be related to flower production. So far this was found to be the case with some developmental stages. DE VRIES (1976a) showed the most productive plants to occur among seedlings with short JP's; in extreme cases such plants may produce three times as many flowers as those with long JP's. In selecting for short JP's the best producers are found, but at the same time selection is made for shorter seedlings. Shorter not only in the seedling stage (DE VRIES, l976a). but also afterwards when the plants are on a rootstock (Fig. 2). Evidently stem length and flower production are opposite characters.

It seems that this dilemma, if recognized at all by breeders, tends to be solved in favour of the long stems, as is demonstrated by the relatively poor production of our modern long stemmed cut rose varieties. Usually stem length and favourable flower properties, or rather in the reverse order, are the main or only characters handled by practical breeders in the first mass selection of rose seedlings.

Because correlations between seedling length and JP, and between production and JP are not perfect, even if high, correlation breakers do occur. It is therefore recommended to select primarily for plants with short JP's, thus ensuring productivity, and secondarily to select in this lot for the exceptions to the rule, i.e. seedlings with long shoots. Using again Ferguson's Table, it follows that in retaining the 50% of plants with short JP's, 20% of them will have an above average stem length.


It is often objected by rose breeders that, where selection for shoot length in young seedlings is still just practicable, selection for JP is not, due to uneven germination of the seeds. This problem can be solved by applying a proper stratification method. When sown in flats and subsequently stored for 3-4 months at 0°C (± 1°) the seeds are ready to germinate. Till that stage actual germination is inhibited by the low temperature, but if the seeds are then placed in a hothouse at 22°C, it is our experience that all viable seeds will germinate within 10 days; the average germination is then about 40%. Each day seedlings in the cotyledon stage can be transplanted and labelled per date. With 10 germination dates there are 10 categories of seedlings. In each category the about 50% of earliest flowering seedlings should be screened for sufficient stem length at 1st flowering. Provided an adequate number of seedlings is available, this procedure increases chances of finding seedlings with desirable qualities (production and stem length) at the moment selection for flower properties is usually done.


The authors wish to express their thanks to Miss Frida Garretsen for constructive criticisms of the statistics.