102 



The subject is so complex and the investigations are so new, 

 that it is unreasonable to expect any approach to finality. Yet it 

 is evident that, no matter what may be said about this or that 

 alleged fact, there is here opened up a most fertile field for research, 

 with reasonable expectations of seeing into the processes of evolu- 

 tion in a manner never before considered possible. Taxonomy, 

 also, takes a new meaning, and may attain more precision than 

 seemed possible in such polymorphic genera. We are encouraged 

 to examine the face of nature, searching for evidence in all direc- 

 tions. New types may come into existance to perish at once, or 

 may survive a little while, and eventually perish, or may spread 

 in ever widening circles. Thus local forms are not necessarily 

 relict species, survivors of a once mighty host. Furthermore, as 

 far as can be seen, there is no reason a species should not be poly- 

 phyletic, in the sense of arising more than once from similar an- 

 cestors, in dififerent places. When this occurs, there will be a 

 probability that the ancestors will not be exactly alike, but will 

 differ in at least some genes, so that the separate colonies (as we 

 call them) of a species may be distinguishable on very close in- 

 spection. That this is true of varieties, we are all aware, and no 

 one doubts that the identical gene mutations occur many times in- 

 dependently. 



It will be interesting to see if novel climatic conditions have 

 any effect on the variable roses. There is a member of the R. 

 canina group supposedly naturalized in Mexico. At Wallangara, 

 on the southern border of Queensland, I found plenty of the ir- 

 regular pentaploid R. rubigirwsa growing. These exotic colonies 

 should be closely watched. If their chance for differentiation 

 depends on hybridization, then they are likely to be more constant 

 than at home. Also at Wallangara, Raimannia odorata, another 

 plant of interest to the geneticists, has run wild. 



It is possible to postulate a scheme of evolution for Rosa which 

 requires neither the breaking up of a hypothetical decaploid nor 

 the building up of polyploids by hybridization. Hurst's di- 

 ploid aggregate "species" may have developed from an ancestral 

 diploid by a process of gene mutations, exactly as the segregated 

 species or subspecies have developed within these aggregates. Then 

 the polyploids may owe their origin to chromosome duplication, 

 and the diversification of their septets (groups of seven chromo- 



