316 PLANTS, SEEDS, AND CURRENTS 



We see it again in the species within a genus ; but in a large genus 

 like Geranium, which holds about 270 known species and has been 

 subdivided by Knuth in his monograph in the Pflanzenreich series 

 into thirty sections, we are restricted by the plan of that work to an 

 appeal to the sections. Small genera are best suited to illustrate 

 the principle, and in my book on the Pacific it is discussed in the case 

 of Dodoncea, Metrosideros, and several others. However, its opera- 

 tion within the sections of Geranium is often clearly exhibited. In 

 such cases the parent species, or the type around which the other 

 species group themselves, is highly variable, and has the largest 

 range, covering the whole or the greater part of the area held by the 

 section. 



Lastly, we see this principle at work within the limits of a species, 

 and it is especially well exhibited in the varieties and local races of 

 some of the species of Geranium. Thus, there are species which, 

 whilst possessing a number of local varieties or races, have a parent 

 form that includes within its range all the homes of the varieties and 

 races. The behaviour of G. mexicanum, as indicated in Knuth's 

 monograph, is very typical. It has the range of the section Mexicana, 

 to which it belongs, all the other species of the section being confined 

 to limited areas. It has a number of local forms, all of which group 

 themselves around a variety that has the range of the species. 



Antiquity and Change. — There is not infrequently an obsession 

 in these matters that time goes with change; and we are now and 

 then apt to look upon some highly differentiated and rare plant- 

 organism as far more ancient than some simple plant-type that 

 abounds around us. This is a dangerous view to hold respecting 

 family types, concerning which it can be contended with much better 

 reason that the contrast is merely a matter of varying rates of differ- 

 entiation. The scores of American genera of flowering plants that 

 have remained unchanged since Cretaceous and early Tertiary times 

 show clearly enough that antiquity by no means connotes change. 

 The only valid explanation of the fact that in one continent a family 

 may be more differentiated than in another, that is to say, that it 

 is farther from the family type, is to be found in the more rapid 

 operation of the process. 



Whilst we should expect to find a primitive family represented on 

 all the larger continental tracts, the absence of some of the derivative 

 families springing from the primitive type is often to be looked for. 

 The differentiation theory takes the world as it is. If we find an 

 explanation of the almost exclusive possession by South America of 

 the Trop&olacea? (one of the Geranial alliance) in the relative isola- 

 tion of that continent, it would surely be inconsistent to postulate, as 

 Knuth does (p. 38), a connection between South America and South 

 Africa in order to account for the occurrence of Geraniacea? in both 

 continents. We must begin with the universal distribution of the 

 primitive parent type of all the families of the alliance, and allow the 

 extent of the differentiation to be determined by the arrangement 

 of the land-masses and their internal conditions. The earlier stages 

 of the process of change would be on similar lines, whilst the later 

 changes would diverge widely. 



