The Constitution of the Rosa Canina Complex

HE peculiar meiosis of the canina roses (sensu lato) was un-T ravelled by TACKHOLM, BLACKBURN and HARRISON, and HURST, in several papers.It was found that the members of this multifarious group possessed 35, more seIdom 28 or 42, chromosomes in their somatic tissue, but irrespective of the chromosome number exactly seven bivalents at meiosis, the rest of the chromosomes forming univalents (7,, f 21,, or 7,, f 14, and 71, $-28,).Trivalents or higher 27This is a regular process.406 AKE GUSTAFSSON and maintained anew by FAGERLrND (1940).Now and then pollen grains with more than seven, and embryo-sacs with less than twentyeight chromosomes are formed, giving rise l o aberrant progeny plants.These are rarely or never met with in natme, however, although not infrequently in experiments.The following problems may be posed: (1) what is actually the method of propagation in the dog-roses, heterogamous fertilization or apomixis?( 2 ) what is the internal constitution of the numerous unpaired chromosomes?and (3) what bearing have the discovered cytogenetical properties on problems relating to the taxonomy of the caninn roses?Finally, (4) we have to investigate how the peculiar type of meiosis once arose, i. e. try to explain the phylogenesis of the group. GENETICS OF THE DOG-ROSES.Up to 1931 it was generally considered that the caninu forms were apomictic, moreover, being autonomously apomictic, needing no pollen for the fruit-and seed-setting.In that year it was shown, however, that wild roses, growing at Svalof and Lund, produced no seeds after emasculation, although hips occasionally developed (CUSTAFSSON, 1931 a, b).The earlier positive results, obtained by numerous Swedish, German and English scientists, did not stand a critical examination.Quite recently FAGERL~NU (1940) was able to confirm this view.In the years 1932, 1934 and 1935 many cross experiments were performed, and these yielded seeds and F , progeny.In the beginning the offspring was thought to have arisen by apomixis (pseudogamy!, inany progenies looking similar to the mother-parent.This idea was however definitely discarded in 1939.The existing similarity is evidently brought about by the prevalence of mother characters.NOW that the plants have fully grown up, the prevalence is less marked.The following hybrid series will be discussed in this paper: R. caninrc var.Blondneana X rugosa (1932-1 and 2, Lund Botanical Garden), R. cnnina I x rugosa (1934-1, Svalof), R. canina I1 X rugosa (1934-2, Svalof), R. rubiginosa X rugosa (1932-5, 1934-2 and 3, Lund and Svalof), R. canina 11 X rubiginosa (1934-5, Svalof), and the reciprocal cross R. rrtbiginoscr X canina I1 (1934-6, Svalof).Summarizing the general results, we must state that the canina roses used in the experiments are one for all sexual'.The progeny This conclusion is at least generally valid, when crossings between different compounds are taken into account.The possibility is not out of discussion, how-~-