Journal of Genetics 57(2-3): 253-268 (July-December 1960)

Gordon Rowley
John Innes Horticultural Institution, Bayfordbury, Hertford, Herts
(Received May 1959)


The chromosomes of Rosa are small and so nearly identical in size and form that they cannot be distinguished individually, but studies on chromosome number provide information of direct value to an understanding of evolution and classification in this difficult genus. All polyploidy levels up to enneaploid (2n=63) have been described. Triploids, heptaploids and enneaploids are known only in a few hybrids and never among wild species.

Aneuploidy has not been reported in wild populations of the genus, and its occurrence is confined exclusively to experimental progenies, most frequently in first and second generation seedlings from crosses between species and between garden roses. The section Caninae provides more than half the known examples, the best source being seedlings raised in cultivation from wild species of Caninae. The occurrence of aneuploidy among experimental progenies and the absence of records on wild populations suggest either a reversion of tissues to the euploid level during seedling development or a competitive elimination of aneuploid forms during the establishment phase from seed.

The present work was undertaken to inquire into this problem.


Täckholm (1920-22) (V. Täckholm 1922) noted variation in chromosome number from 2n=34-40 in some F2 and later generations of hybrid Caninae. The examples quoted were the following:—

  Observed Expectation
R. alba x canina 34 35
R. x pokornyana (= canina x rubrifolia) 40 35 or 42
R. coriifolia matssonii x corrifolia incrassata 36 35
R. coriifolia x glauca 37 35

Heslop Harrison and Blackburn (1921-55) (v. Harrison 1955) reported aneuploids with "2n=24, 2n=32 and so on", but no further details have been given in print yet. Erlanson (1929-33) described seedlings of Rosa blanda (2n=14) with 15 and 16 chromosomes, and one of diploid R. pyrifera with 2n=16. Six aneuploids were found in diploid and tetraploid species, one family giving the series 2n=14, 15, 16 and 17. In two of the aneuploids, R. pyrifera (2n=16) and R. arkansana (2n=31) the extra chromosomes could be identified owing to their smaller size. Gustafsson (1944) recorded seedlings with 2n=34 from R. canina♀ x rugosa♂ and R. canina♀ x rubiginosa.♂ No aneuploid counts appear among the very large number of determinations made by Dr. C. C. Hurst (unpublished), although he was aware of the existence of aneuploids and listed those found by Täckholm. Hurst's interest was mainly on the euploids as supporting his Septet Theory in which the chromosomes were assumed to act only in blocks of seven.

Raby (1937-38) and Sansome, Raby, and Fothergill (1939) found aneuploids among the strains of Caninae used as understocks. This is rather unexpected, since these strains are selected for maximum fitness and vigour. 'Kokulensky' was recorded as having 2n=36 and 'Senff' 2n=34; the latter raised from seed at the John Innes Horticultural Institution has shown the same aneuploid number in nearly all its progeny. Also of interest is the case of "Rosa laxa var. microcarpa", where seedlings were illustrated with 2n=14, 15 and 24. Unfortunately, further study of these unique plants is impossible as neither name nor plants can now be traced.

Of more than a hundred known cases of aneuploidy, seven-eighths have been detected by Miss A. P. Wylie working on the National Rose Species Collection at Bayfordbury. They are summarised in Table 1; all but five are published here for the first time.

The frequency of aneuploidy in the genus can be assessed from Fig. 1. Ths covers every ascertainable published and unpublished aneuploid count. Aneuploids are most frequent at the pentaploid level, the dominant ploidy of the Section Caninae. At the diploid and triploid levels the paucity of reported aneuploids is presumably due to the severe unbalance from loss or gain of a chromosome as compared with higher polyploids. The frequency of distribution of aneuploids around the higher euploid levels in Fig. 1 is fairly symmetrical. The apparent asymmetry around 4x=28 could well be fortuitous or due to subjective interpretations in chromosome counting. The latest counts given Table 1 show approximately equal distribution around each euploid level.

Aneuploids are most frequent in the Caninae Section in which the mechanism of chromosome separation at meiosis is unique. Here 14 to 28 chromosomes are transmitted unpaired through the female gamete while the male gamete contains only 7. Whether or not this system of reproduction, where a major part of the chromosome complement does not undergo recombination, is responsible for the prevalence of aneuploidy in the section cannot be decided with certainty. The influence of this sub-sexual type of chromosome mechanism on the effect of loss or gain of chromosomes in aneuploids cannot at present be distinguished from the effects of polyploidy per se, where changes in the chromosome complement are buffered by the presence of several homologous chromosomes.

In the field, Caninae roses show different types of behaviour in different localities. Over large stretches of country a single phenotype of a species will be found to dominate with no appreciable variation. Likewise sowings of wild seed often produce progeny so uniform that earlier observers concluded that they must all be apomicts. The Caninae used for rootstocks mostly have this desirable trait. On the other hand there are areas where no two rose bushes seem exactly alike, showing differences in the disk shape, leaf serration, armature density and other morphological as well as physiological and biochemical variations. A similar diversity sometimes, though not often, manifests itself among cultivated seedling Caninae.

In the analyses of the families of Caninae seedlings presented here morphological variation with and without chromosome variability is described. No case has been found of aneuploidy in phenotypically uniform families except in one family of Rosa stylosa where one tetraploid plant was found which was indistinguishable from the normal pentaploid form except for reduced fertility.

Table 1. New Aneuploid Rose Counts1

I. CANINAE   30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 63
R. afzeliana           8 8 2                  
R. canina   2   4 3 15 3                    
 "       "  'Grange Briar'                   1     7 2      
 "       "  'Kokulensky'                                  
 "       "  'Senif'         1 9                      
R. canina X arvensis         9 0                      
 "       " X rugosa) F2         1 *             1        
R. coriifolia                 * * * * * * *    
 "       "          2 10 4 3 1                
R. elliptica           *                 1    
R. x involuta wilsonii 1 1 1 5 1 1     1   1            
R. jundzillii                         * 1      
R. sherardii pseudomollis         1 *                      
R. stylosa         1 4   1                  
 "       "  systyla         1 *                      
 "       "  X canina           2 1                    
R. sp. (Geneva)                   2              
R. sp. (Spain)                       1 *        
II. PIMPINELLIFOLIAE   27 28 29                                  
R. spinosissima dunwichensis   * 1                                  
 "       "  hispida   3 1                                  
III. CINNAMOMEAE 21 22                                        
R. X kochneana 1                     1                    
R. pendulina oxyodon *                     1                    
(R. pendulina X moyesii) F2           1         1 1 1     1           1
IV. GARDEN ROSES                                            
'Amurensis' seedlings       1               1                  
'Bonn'       *                                  
‘Eos'                           1                
'Poulsen's Crimson' *                                        
'Refulgence'                                 *        
'Rosenwunder'       *                                  
'Tapis Rose'     *                                    
* Number unspecified
1 All counts listed in this table were made by Miss A. F. Wylie.


Six families have been studied comprising three wild species, Rosa afzeliana (glauca), R. canina and R. coriifolia; a wild F1 hybrid R. x involuta wilsonii and two Caninae selections used as clonal stocks: 'Grange Briar' and 'Senif'. Each family was originally represented by 20-27 individuals raised in a closely planted block to allow of a certain degree of competition between the plants. Morphological characters were scored at flowering and fruiting time using the score sheets adopted by the British Rose Survey in which 36 of the most useful taxonomic characters are tabulated with series of 3 to 11 contrasted descriptions. Only brief summaries of these scores are presented here.


1. Rosa afzeliana Fries (R. glauca Vill.). H 622

Source: Open-pollinated seed collected in Middleton-in-Teesdale, Sept. 1949 by Mrs. R. Hurst in an area where the rose populations showed great diversity. The parent bush was described as having more or less paired prickles, slender fruits and erect to spreading pinnate sepals.

Taxonomy: This family showed considerable variation in all characters scored except for two: the slightly protruding styles and the lack of tomentum on the leaflets. This latter feature separates the species from the closely allied R. coriifolia. Three individuals had a moderate development of glands on the rachis, sufficient in one case to give noticeable sweet-briar scent. Six plants were of very vigorous habit, two remained dwarf and one was a "runt".

As determined by WoIley-Dod's (1930) keys, the plants could be classified into four different "varieties", two of which have even been considered as separate species by earlier botanists:

A Sepals erect or inclined on fruit, persisting until it is ripe.

R. a. glaucophylla (Winch) W.-Dod

(R. glaucophylla Winch)

2n=35 (1 plant), 36 (1 plant)