OF THE DILLENIACE^. ■- 75 



this to the Dilleniacese, we get a tree somewhat Uke the 

 following (of course these are my own views as to the relation- 

 ships, and other people will have different views) : — 





^TetracerM 



The order of the names in the top row does not, of course, 

 indicate nearness of relationship. The diagram is an imaginary 

 history of the mutations, and Delima, e.g., is often united to 

 Tetracera. 



If now, on this basis, we draw a map of the evolution of this 

 order and its distribution , and assume the former existence of 

 the continent, Brazilia-ethiopica, with land connections to 

 Ceylon and Malaya, we get a surprisingly simple picture, 

 showing that the order probably originated with Tetracera 

 and Saurauia, and of these the former, being the simpler, 

 was probably the older. It would appear to have started in 

 Brazilia-ethiopica, and spread thence to South America, Africa, 

 Ceylon, and Malaya, and from the last to Australia. We are 

 no longer to be led astray by the great numbers of species in 

 Malaya- Australia , for example, for we have learnt that, as a 

 rule, local species do not count. Evidently when it reached 

 Australia the order proved very suitable to that country, for 

 there are now great numbers of Hibbertias, &c., there. 



