THE DISTRIBUTION OF COMMUNITIES 353 



relict communities. They are ill adapted for combat with the ubiqui- 

 tous immigrants brought in everywhere by cultivation. 



The rock-crevice associations of the headlands of the Mediterranean 

 coasts from the Pillars of Hercules to Crete and Cyprus are relict 

 associations of Tertiary origin. They are due to edaphic conditions 

 and are rich in remarkable endemics. Maire and Braun-Blanquet 

 have described a very local edaphic relict association in the Andryale- 

 tum mogadorensis of the island of Mogador. As the island is in process 

 of destruction by the waves of the sea, the fate of this association is 

 sealed. The forests of Pinus silvestris described by Schmid (1929), 

 strictly limited by the rocky and gravelly substratum of limestone 

 and dolomite, constitute relict associations (see also Braun-Blanquet, 

 1932). 



The question of whether a given association is a pioneer or a relict 

 acquires a special interest on the boundary of a climax region. The 

 answer may indicate probable changes of climate. Are the forest 

 islands in the alpine levels, or in the Podolian steppes, and the steppe 

 associations in Bohemia and Moravia to be regarded as pioneers or 

 relicts? When this is decided, we can calculate the extension of one 

 climax region and the reduction of the other. 



The difficulties which attend the answering of such questions, 

 especially where human influences have disfigured the original vegeta- 

 tion till it is unrecognizable, are shown by the endless differences of 

 opinion among North American investigators as to the natural bounda- 

 ries of forest climax and prairie in the midwestern states. 



There is a special group of very striking relict associations which 

 at present are found only on artificial habitats of human origin. They 

 have no known natural habitat. The elegant Centunculo-Anthocere- 

 tum punctati of clayey, non-calcareous stubble fields, with Delia 

 segetalis, Montia minor and Myosurus has no natural habitat, in 

 Switzerland or in the upper Rhine country (Koch, 1926, p. 25). Also 

 the well-characterized miniature association of Cicendia and Stereodon 

 arcuatus, which Allorge (1922) and Gaume (1924) have described from 

 little-traveled wood roads in western and northwestern France, seems 

 to occur only very rarely in natural habitats. 



Vegetation Territories. — Synchorologic synthesis culminates in 

 the definition and description of natural vegetation territories. This 

 geographic concept has arisen from the related idea of floristic regions, 

 rather independently of the findings of plant sociology (rf. Braun- 

 Blanquet, 1919). 



After the first attempts of Willdenow (1797), Treviranus (1803), 

 and Humboldt (1805, 1807), the Danish scholar Schouw (1823) and 

 the Swiss Auguste Pyramus de Candolle (1820) were leaders in estab- 



