American Rose Annual (1956) 41: 70-73
Germination In Rosa Canina
John Innes Horticultural Institution,
Bayfordbury, Hertford, England

Roses vary widely in the ease with which the seeds germinate. In some, as the species of the Section SYNSTYLAE, germination may exceed 90% within a few weeks of sowing. Rosa multiflora is a good example. On the other hand, the Dog Roses (CANINAE) have mostly low viability, often not more than a third of the seeds producing seedlings, and these erratically over a number of years. Typical of this delayed germination is Rosa canina, as shown by the following figures for 600 seeds sown in the autumn of 1949. Seedlings were pricked out as they came up and the seedpans kept for six years.

Years after sowing Percentage Germination
1 0.7  
2 30.7  
3 1.2  
4 0.7  
5 0.3  

This type of behaviour is no doubt a boon in nature, where it enables the species to survive after adverse seasons, but it is a bugbear to the nurseryman and breeder, for whom a year's delay means a waste of space and labour.

Experiments have therefore been carried out at the John Innes Horticultural Institution over the past seven years in an effort to speed germination in Rosa canina.3 This species was chosen because of its wide popularity as a stock — a stock which, incidentally, is as awkward to propagate vegetatively as it is from seed. The findings have indicated the optimum conditions for harvesting and storing seeds, and have led to a technique which can be applied with varying amounts of success to other species of Rosa and to hybrid seed.

Figure 1: Achene of Rosa canina split longitudinally, showing thick wall, bristles and central cavity.
Figure 2: The same cut transversely, showing single seed cut across the cotyledons.
Figure 3: Seed removed from achene.
Figure 4: The same with testa removed showing small embryo and two large cotyledons.
Figure 5: Germinating achene.

First let us look at the seed of the Dog Rose itself — or, rather, what is popularly called the "seed" (Figs. 1-5). To the botanist it is a one-seeded fruit or achene, very much like the nut of a hazel in miniature. The true seed (Fig. 3), with its brown leathery skin or testa, lies snugly within the very hard, bony achene, which usually has a tuft of bristles at the end to which the style was originally attached. For germination to take place, water has to penetrate first the achene wall and then the testa, which bursts apart with the expansion of the first root (Fig. 5).

The causes of delayed germination in plants have been reviewed in detail by Barton and Crocker.2 They include mechanical barriers to water absorption such as an impermeable seed-coat and physiological barriers within the seed necessitating a period of after-ripening before it can reawaken. Authorities differ in their reports on the behaviour of Rosa and it seems likely that both types of barrier operate in different amounts among different species and hybrids.

Figure 6: Comparison of germination in new and old seeds of Rosa canina. Seeds were harvested in the autumn and sown at once. The old seed, which had after-ripened naturally on the plant, germinated the first spring after sowing.

Embryo culture has been successfully applied to roses for the raising of small amounts of seed,1 but it is too laborious for the mass raising of stocks commercially. The methods described below were aimed at speeding the after-ripening by finding the optimum time of collection and storage conditions.

Extensive collection of hips from a single R. canina bush first revealed how the set of achenes varies from year to year under natural conditions of pollination. Samples for 1947 averaged 26 achenes per hip, and on dissection these proved to contain 94% plump seeds. The following year the average was only 16 achenes per hip, of which 88% contained good seeds. When thrown into water and stirred briskly to avoid trapped air bubbles, all the achenes that float are found to be empty and hence inviable. However, a few that sink have shrunken or malformed contents and would never germinate. Hence the flotation test is not an absolute guide to the total of good seeds. Among the achenes collected, one always finds a number of miniature size but with apparently plump contents: trials of these failed to germinate any of them.

Figure 7: Comparison of different methods of storing stratified seeds of Rose canina. Each batch was cleaned from the hips and kept at the first temperature for two months, transferred to the second for two months, and then sown.
C = Cold M = Medium H = Hot
A high followed by a low temperature consistently gave the best results.

Germination the first year was found to improve if the achenes were stored in a moist medium as compared with storing dry or in the hips (Fig. 8). Attention was then paid to the temperature during storage, and in this respect a constant temperature proved less effective than a change from one extreme to another during storage. Of various combinations tried, a high temperature for two months followed by chilling for the same period gave the best results (Fig. 7). Various media were tried for storing the achenes, such as the traditional peat and sand and mixtures of the two. But far better results came from using horticultural vermiculite ("Exflor," coarse grade) because of its light, open texture, good aeration and moisture retention.

Figure 8: Comparison of germination in Rosa canina when seeds are stratified after cleaning (back row) and stratified in the hips (front row). Three different storage methods were also compared (c.f. Fig. 7).

At one time it was thought that the hip flesh itself exerted a retarding influence on the contained seeds, so the experiment was made of enclosing achenes of proved good germination (Rosa carolina, helenae, multiflora, omeiensis, rugosa and spinosissima) in emptied canina hips and comparing their spring germination with controls stored in their own hips. However, the two batches showed no significant differences. Nor did achenes from terminal hips of R. canina germinate in any way differently from those in lateral hips, in spite of often striking differences in the size, shape and set of lateral hips.

However, one factor was found to exert a great influence on both total viability and speed of germination. That is the time at which the hips are harvested. Preliminary tests showed that prematurely harvested seeds (that is, those collected at any time between petal fall and while the hips are still green and the sepals fresh) gave negligible germination over four years (only 1 in 300). Figures for batches of seed collected at intervals from the same R. canina over the later stages of ripening are as follows: —

1950 Seed
Date of Hip Collection Hip Colour and Texture Condition of Sepals Percentage Germination
1951 1952 1953
Sept. 8 Flushed orange, firm Withered but persistent 0 17 0
Sept. 22 Red, firm Falling 0 53 0
Oct. 6 Red, firm Fallen 0 60 0
Oct. 20 Red, firm Fallen 1 14 0
Dec. 1 Red, soft and jammy Fallen 0 13 0

In the above trials the achenes were sown directly after harvesting. On this evidence the optimum time for harvesting would seem to be as soon as the hips are fully ripe but before the flesh has begun to soften. However, individual Dog Roses show variation among themselves, both in the time of sepal fall and the total viability, so one cannot generalize from the results. Some strains of R. canina are less attractive to birds than others, so that a number of hips stay on the plant for over a year, withered and dry. Seeds from these hips, which have after-ripened naturally on the bush, germinate at once (Fig. 6).

By piecing together the findings of these and other experiments it has been possible to formulate a general procedure by which seeds of Rosa canina can be persuaded to skip a year and germinate as well the first spring as it would, untreated, the second. This procedure has been adopted as a matter of course for all rose seed, hybrid or otherwise, at the John Innes Institution. The hips are harvested as soon as ripe in the autumn, opened without delay and the achenes cleaned out. They are stratified in tall pots ("long toms") of moist vermiculite, screened through an appropriate sieve so that the seeds can be sifted free with ease in the spring. Two or three layers can be included in one pot. The pots are then stood in a warm glasshouse for two months and then transferred to a refrigerator at approximately freezing point for a further two months. Care is taken to dampen the vermiculite so that at no time do they dry out completely. Finally the achenes are sifted out and sown in the normal way.

A simpler method, if no refrigerator is available, is to bed the pots in ashes outdoors during the winter. It gives almost as good results, but protection from mice and rats is necessary.

1. ASEN, S. & LARSON, R. E. Artificial culturing of rose embryos. Pennsylvania State College Progress Rep. No. 40 (1951). 4 pp.
2. BARTON, L. V. & CROCKER, W. Twenty Years of Seed Research. London 1948.
3. ROWLEY, G. D. in John Innes Hort. Inst. 49th Ann. Rep. (1953) 27-8.
4. TINCKER, M. A. H. Rose seeds: their after-ripening and germination. J. Roy. Hort. Soc. LX (1935) 399-417.