Evolution 13: 378-388. (Sept 1959)
Department of Genetics, University of Warsaw
Received November 17, 1958


The genus Geum of the family Rosaceae consists of about 56 species widely spread over the world except in the tropics. All species are perennial herbs growing in more or less mesic conditions. More than fifteen years' study of this genus by the present author have shown that this genus is a polyploid complex of which the various species represent different stages of evolution, and show interesting relationships.

In spite of great morphological differentiation on different polyploid levels the crossability between different species is surprisingly high, permitting cytogenetic analysis of numerous interspecific hybrids. Geum presents many interesting general problems concerning the role of geographical and ecological isolation as well as hybridization and polyploidy in the process of speciation. Some of the more interesting results of published (Gajewski, 1948, 1949, 1950, 1952, 1953, 1955, 1957) as well as unpublished results will be presented here.


The species of Geum are rather well characterized and are easy to separate except for a few recent groups that show intensive hybridization or introgression, chiefly in areas where the original vegetation has been disturbed by man. On the other hand, the separation of Geum from related genera has been much discussed among taxonomists without any general agreement. The results of cytogenetic analysis have clarified to some extent this homologous, it seems more reasonable to species is an old polyploid complex in which different genome combinations represent distinct groups of more or less related species. The separate groups often differ very conspicuously in morphology and geographical distribution. They have probably arisen in different times. This was the cause of the separation of several groups of species in as many as six or even eight different genera (Rydberg, 1898, 1908-1916; Bolle, 1933; Jezupchuk, 1941). As the hybrids between some of these groups are still partially fertile and some of their genomes are homologous, it seems more reasonable to keep them together in one genus subdivided in many subgenera. This classification, I think, gives a better insight into the genetical relationships of the taxa studied, whereas splitting them into numerous genera obscures phylogenetic relationships.

The genus Geum (sensu lato), with the closely related but clearly separate genera Waldsteinia and Coluria, belongs to the tribe Geeae. This tribe, with the tribes Cercocarpeae (genus Cercocarpus) and Dryadeae (genera Purshia, Cowania, Fallugia and Dryas), forms a separate subfamily, Dryadoideae of the rose family. This subfamily is clearly separated from the related subfamily Rosoideae by having orthotropous ovules, and represents a separate evolutionary line.

FIG. 1. Diagram showing the relationships of the genera and subgenera in the subfamily Dryadoideae. The dark patches represent approximately the relative size of the different taxa. The outer circles show basic chromosome numbers and the type of the structure of the styles.

The origin of the subfamily Dryadoideae is probably so old that the first steps of its evolutionary history can only be guessed. It seems very probable that the woody genera of the tribes Cercocarpeae and Dryadeae (except perhaps Dryas) confined to Western North America are the present day survivors of an old primitive group of Dryadoideae. They show now, of course, many derived characters acquired during their long evolution, as for instance the reduction of the number of pistils to one per flower in Cercocarpus. However, their restricted distribution, woody habit, long persistent and pennate styles adapted to wind dispersal seem to be more primitive than the herbaceous habit and styles specialized for transport by animals which are characteristic in most Geeae. All species of these two tribes, including Dryas, have the basic chromosome number x = 9 and only diploid species are known, whereas in the widespread tribe of Geeae with only herbaceous plants the basic chromosome number is 7 and polyploids are very numerous. A case of very curious parallel evolution seems to have occurred in the related subfamily Rosoideae with anatropous ovules. The tribe Kerriaeae with few, mostly monotypic, genera geographically restricted to the East and West shores of the Pacific like Lutkea, Gillenia, Neillia, Pentactina or Stephanandra, has also the basic chromosome number 9 with only diploid species known. The other tribes of this subfamily like Roseae, Potentilleae or Sanguisorbeae are mostly herbaceous with the basic chromosome number x = 7, and most genera are widespread in the Northern Hemisphere with numerous polyploid species. It seems thus that the tribes in both subfamilies with the basic chromosome number 7 are the youngest. Why the evolution from the shrubby to the herbaceous habit was often connected with the change of the basic number from 9 to 7, and why the 7-chromosome genera have wider distribution, have formed numerous species, and show numerous polyploids is now difficult to say. Some possibilities to explain these phenomena will be perhaps evident from the results of the work on Geum. In any case it is evident that these plants are younger. Many of their numerous amphiploids grow on glaciated areas, and their origin was probably influenced by Pleistocene migrations and partly also in most recent times by man. Some of the relationships among different taxa of the tribe Geeae are shown in figure 1.


Chromosome Changes in Evolution and Adaptation