176 H. B. GUPPY, M.B., F.E.S.E., ON PLANT-DISTEIBDTION 



with which have been disguised by later modifications. To raise 

 a specialised genus to the dignity of a family is to reverse the 

 natural order of things. It will be pointed out later on that 

 nature seems to have reversed the regular process of differentia- 

 tion in Oceanic islands, and in other localities where conditions 

 of abnormal isolation prevail : but it will be shown that it is 

 not nature tliat has reversed her processes, but the botanist 

 that has changed his methods. If we accept the single-centre 

 hypothesis, the birth of families presents itself as a very 

 haphazard operation. One ranges the world, whilst another is 

 confined to tlie tropics, a third to the north and south temperate 

 regions, a fourth to only one of tlie temperate zones, a fifth to 

 North America aud Eurasia, a sixth to one continent only, and 

 so on. By regarding these matters from the standpoint of the 

 differentiation theory, we shall see that the difficulties have 

 been largely created by ourselves, more especially through the 

 loose employment of the term " family."* 



We have first a world-ranging family, like the Compositae, 

 where the various tril)es, differentiating in situ, collectively 

 occupy tlie area of tlie })rimitive ty})e. In other cases, however, 

 differentiation has proceeded so far that the original tribes are 

 ranked as families by the systematist : and we obtain a series of 

 related families, each in its own region, l)ut together holding 

 much of the area of the world-ranging original type. Thus the 

 closely related families of the PrimulaceiTe and the Myrsinaceie, 

 the first of the tem})erate regions, the second of the tropics, 

 may represent the tribes that indicated the first step in the 

 differentiation of a parent type that was once generally 

 distributed over tlie earth when climatic conditions were more 

 uniform than tliey are at present. Just as we may regard a 

 widely distributed laruily like the CompositaB as representing 

 in its tribes, genera, and species the history of the differentia- 

 tion of flowering plants and of their conditions of existence on 

 the globe, so we may see in tlie closely related Myrsinaceie and 

 rrimulaceic and in their respective genera and species the 

 result of the differentiation of plant-l'orms and plant-conditions 

 since the era of flowering plants began. 



The custom among systematists of linking familits together 

 in such a way as to suggest a genetic connection offers evidence 

 in favour of the difierentiation hypothesis, more especially in 



* Faniihes should be ranked in grades according to their relation to 

 tlie parent type. So also as regards genera and species the same system 

 slionld be u.sed. 



