September 15, 192 1] 



ffATURE 



99 



British Roses and Hybridity. 



THE genus Rosa, along with the brambles (Rubus), 

 hawkweeds (Hieracium), and certain other poly- 



norphic genera, has long furnished serious problems 

 :o the systematic botanist. Darwin spoke of the kind 

 of variability which these genera show as ' indepen- 

 dent of the conditions of life," meaning that the 

 differences exhibited are "of no ser\ice or disservice 

 to the species," and are therefore not directly subject 

 to natural selection. Modern geneticists would largely 

 agree with that point of view. 



In two recent papers (^Annals of Botany, vol. 35, 

 p. 159, and Trans. Nat. Hist. Soc. Northumberland, 

 Durham, and Newcastle, vol. 5, part 2) Dr. Harrison 

 and Miss Blackburn have thrown a great deal of light 

 on the polymorphic condition which has long puzzled 

 students of British roses. Their results are based cmi 

 a careful study of the roses of northern England in 

 the field and a cytological investigation of their 

 chromosomes. This type of combined cytological and 

 experimental investigation has already elucidated the 

 nature of the variability in such genera as CEnothera 

 and Drosophila, and its application to other groups is 

 probablv the most promising field in present-day genetics. 

 Contrary to the early studies of Strasburger, the 

 fundamental haploid number of chromosomes in the 

 genus Rosa is found to be seven. In Rosa aroensis 

 and related forms the diploid (2X) number of chromo- 

 somes is fourteen. These forms show normal fer- 

 tilisation, and the meiotic divisions are free from 

 irregularities. The group of forms to which R. pim- 

 pinellifolia, L., belongs is tetraploid, the somatic 

 number of chromosomes being twenr\-eight (4X). In 

 these forms also the meiotic divisions are for the 

 most part normal, and fertilisation is necessary. 

 These two groups of roses form a parallel to the 

 diploid and tetraploid types in species of QEnothera, 

 and since tetraploidy is known to have originated by 

 mutation in CEnothera, it has probably originated in 

 the same manner in the roses. The group of roses 

 known as \'illosae is also tetraploid, but that these 

 roses have originated in another way, probably as 

 hybrids, is shown by the behaviour of their chromo- 

 "omes, for instead of forming fourteen pairs in the 



-duction division there are seven pairs and fourteen 

 -.ngle chromosomes. 



Perhaps the most interesting are the pentaploid 

 v5x) roses, of which four grouos are recognised, con- 

 taining such species as R. sylvestris, R. rubiginosa, 

 and R. coriifolia. These have thirty-five somatic 

 chromosomes and twenty-eight in the reduction divi- 

 sion. Of the latter bodies seven are pairs or bivalents 

 and twent}--one singles or univalents, thus making up 

 the full number. The bivalents are almost invariably 

 arranged in a central group surrounded by the uni- 

 valents. These pentaploid roses hc've probably all 

 originated through crossing, as is indicated by the 

 behaviour of their chromosomes, among other things. 

 The reduction divisions following this manner of 

 chromosome pairing show the familiar irregularities 

 in the distribution of their chromosomes, with the 

 result that the pollen is almost wholly sterile. Dr. 

 Harrison has also shown experimentally that these 

 forms will set seed apogamously, although producing 

 fertilised seeds in the same flower. The presence of 

 apomixis, of course, accounts for these roses ccKning 

 true from seeds. 



The diploid roses and the Pimpinellifoliae, which are 

 sexual, are, on the contran,', looked upon as pure 

 species. In accordance with the views of Ernst in a 

 book on apogamy, published during the war, the 

 presence of apogamy is regarded as a result of the 

 stimulus of hybridity to vegetative development. The 



NO. 2707, VOL. 108] 



authors would explain the origin of many British 

 nybrid roses of this type through northern forms be- 

 longing to the Afzelianae clashing with the southern 

 Eucaninae, owing to climatic changes occurring per- 

 haps at the beginning of the Glacial period. Other 

 hybrids are probably being formed even now, and 

 since the Afzelianae are themselves pentaploid (5X) 

 they must be the product of still earlier crossing. 

 Thus it seems clear that a large proportion of our 

 rose species are ancient hybrids the characters of 

 which are perpetuated by apomictic reproduction. 



Mr. J. R. Matthews (Ne-d.' PhytoL, 1920, p. 153) 

 has also made a systematic study of British rose 

 hybrids, and Miss Co'e (Bot. Gazette, 1917) has 

 examined the pollen of American roses. 



These and similar results obtained in recent years 

 raise several important questions. The first is, of 

 course, the part plaved by crossing in connection with 

 evolution. The most extreme views on this subject 

 have probably been held by Lotsy and Jeffrey. Both 

 agree in believing that hybridisation occurs practically 

 throughout the Angiosperms. The former, however, 

 would attempt to explain all evolution as the result 

 of crossing, while the latter holds that hybridisation 

 and hvbrid forms can have played no part whatever in 

 evolution. In this connection it seems well to remem- 

 ber that in open-pollinated plants the evoluticwiary 

 unit will be a group of interbreeding forms differing 

 from each other in various characters, the differences 

 being perpetuated bv the crosses between individuals 

 which are made each year. Crossing between related 

 forms is then a condition under which evolution has 

 taken place not only in open-pollinated plants, but 

 also in animals. It does not necessarily follow, how- 

 ever, that crossing is the cause of that evolution. Dr. 

 Harrison and Miss Blackburn conclude that " hybridiU' 

 is one of the prime factcM-s in the evolution of species," 

 if not the only one. Their own results with the tetra- 

 ploid (4 X ) roses of the group Pimpinellifoliae, however, 

 Indicate that after the tetraploidy originated (whether 

 through the stimulus resulting from crossing or not), 

 germinal changes, i.e. mutations, must have occurred, 

 giving rise to the present group of tetraploid forms. 

 Again, the occurrence in various rose species of micro- 

 genes differing from each other in having full green 

 or glaucous foliage, glabrous or hairy leaves, etc., can 

 probably be best interpreted as the result of what we 

 now call parallel mutations rather than indicating 

 orthogenesis, as Dr. Harrison suggests. 



Viewing these and related facts as a whole, it 

 seems clear that while mutation and crossing are 

 almost inextricably intermingled in the history of many 

 plant genera, yet it does not follow that crossing is 

 the cause of mutation. Even if this were the case, 

 they should still be regarded as separate phenomena. 

 It is a curious fact that in Drosophila, where the sexes 

 are separate and crossing of individuals must there- 

 fore take place in ever\- generation, no one seems to 

 have tried seriously to explain the mutations as a 

 phenomenon merely of hybridisation. This is probablv 

 because the evidence apoears strong that manv of 

 these mutations had their germinal origin at about 

 the time they appeared as external characters, i.e. in 

 the immediatelv preceding generation of germ-cells. 



Another point of general interest in this connection 

 relates to the pollen sterility of hybrids. JeTfrey and 

 his pupils have claimed that the presence of "bad 

 pollen " is a proof of hybridit>-. This is the revival 

 of a much older view, and studies of the pollen of 

 Epilobium, CEnothera, Rubus, and other genera have 

 been cited in support of this contention. Some of the 

 results have been severely criticised bv svstematists. 



