CH. x] AGE AND AREA THEORY 103 



cerned with precisely the same point. The question may be im- 

 answerable; but there are those who might see in a primeval 

 jumble of family types the backgroimd of the Avhole story of 

 Distribution. They might regard it as the most significant indica- 

 tion of the great antiquity of the higher plants, and they would 

 see in this world-wide mixture of family types the impress of the 

 lost Mesozoic ages on the history of the flowering plants, ages of 

 unceasing revolutionary changes in the relations of land and sea. 

 They would see in this world-spread mixture the materials on 

 which the great laws of de\'elopment ha\^e operated in the later 

 ages. Such would be their standpoint. But the problem may 

 prove to be one for the biometrician; and we may perhaps be 

 able to learn from him in the case of other world-spread mixtures 

 of organisms of different types the significance of the de\elop- 

 ment of uniform mixtures of types in Time. 



There is another way of approaching the central problem of 

 Distribution, and that is best typified in the case of the gold- 

 miner who, guided at first by a faint show of colour in his pan, 

 follows the clue through until he finds the reef. This is pretty 

 much what Dr Willis has been doing for years in the working out 

 of his Age and Area theory. With a history of small beginnings in 

 Ceylon long ago, it is still in the making, and we can watch its 

 development. It is assimilating as it grows numbers of ideas that 

 have been floating in the minds of biologists for generations, and 

 linking together others that ha\e alwa3's been diflicult to place. 

 Its tendency to luiify and co-ordinate as it develops are two of 

 its striking features. The writer's attitude towards it may be 

 thus stated. Recognising that we had here a courageous and 

 persistent effort to utilise the statistical method in getting behind 

 the distribution of living plants, the question whether it was 

 wrong in this or wrong in that did not seem to be of primary 

 importance. For years the writer had been approaching the 

 subject of Distribution from the opposite direction, that is, from 

 the a priori side. Like many a general theory that had not been 

 linked up A\nth the other side the one that he advocated (a theory 

 of differentiation of generalised types) stood still for lack of 

 verification; and there were echoes in his memory of the despair- 

 ing counsels of those in this and other lands who regarded Dis- 

 tribution as beyond the pale of human endeavour. So that when 

 he realised the possibihties of far greater extension that lay 

 behind the Age and Area hypothesis, the question for him was 

 not whether Dr Willis was right or whether he was wrong, but 



