CHAP. LXiv FORMATIONS ON SHINGLE AND RUBBLE 247 



arc necessarily to some extent different from those on stationary talus. 

 On the former, the proper fixation of the plant is a matter of difficulty ; 

 the vegetation is constantly destroyed by movements of the rubble 

 and must start afresh. The particularly successful species are largely 

 those possessing very long roots or long creeping rhizomes.^ This kind 

 of substratum is poor in plants and the vegetation remains open, until 

 the soil is stationary or ceases to receive fresh masses of stones arising 

 by weathering of the rocks. 



These formations are closely allied with those of fell-fields, witli 

 whicli they merge. In regard to the number of formations and associa- 

 tions on rocks future research must dccide.- 



' SceC. Schroter, 1904-8. 



* Special reference should be made to the papers by Hitchcock (1898), Ottli 

 {1903), Pohle (1903), Rikli (1903), Adamovicz (1898), Ostenfeld (1906), Brockmann 

 Jerosch {1907). For the most recent treatment of the subject of formations and 

 associations, see C. Schroter, 1904-8; also C. Flahault, 19066. 



