CH. xxii] GEOGRAPHICAL DISTRIBUTION 243 



plan, new mutations being cast on the average at a fairly definite 

 speed, differing of course for different classes of animals and 

 plants. 



The acceptance of the view that B is the direct descendant of 

 A, another living species, instead of both being the descendants 

 of some hypothetical a (an ancestor which by the way has never 

 been found in the fossil state, so far as I know, though on the 

 current theory there should be hundreds of thousands of them) 

 will make the work of tracing phylogenies easier, though if 

 mutations may be of large size, this will not always be easy. 



Except in cases where we have geological evidence of former 

 greater spread, when of course the "fossil" area must be added 

 to that occupied by the living plants, we may leave out of 

 accovmt the more local genera in tracing phylogenies, and it is 

 clear that species or genera that are widely separated in space, 

 and in whose case no fossils can be found filling up the spacial 

 gap, cannot, without great risk of error, be looked upon as 

 necessarily closely related, however much alike they may be 

 (cf. 130, p. 346). Their resemblance may be due to parallel 

 mutation, and their ancestors may have been more widely 

 separated than they themselves are. 



In the same way, no fossil that is not of wide dispersal can 

 safely be regarded as an immediate ancestor for anything that 

 is of equal or nearly equal dispersal, and still less if it be of 

 greater. Nor must widely separated fossils be regarded as nearly 

 related Avithout links. Nor is it safe to regard two layers as of 

 the same horizon without a number of fossils in common; and 



Age and Area also throws light upon the question of Floral 

 Regions, which are usually defined as marked out by containing 

 large numbers and proportions of endemic forms, and as being 

 the better marked and more natural the higher the rank of these. 

 Great difficulty has always been encountered in defining such 

 regions; and to make them agree with those of the zoologists is 

 usually regarded as hopeless. In the accepted grouping of them, 

 the southern regions are very much smaller than the northern, 

 owing to the fact that endemics increase in number and propor- 

 tion towards the south (p. 149). Thus south-west South Africa 

 is regarded as a region equivalent to the Mediterranean region, 

 which includes all the land around that sea as far as Beluehistan; 

 and even to the whole of tropical America, including the ^Vcst 



16—2 



