EVOLUTION OF ECOLOGICAL CONCEPTS 345 



vegetation have been very widely observed; but the same recent studies 

 of the eastern forests support Ramensky's view that vegetational continuity 

 is the more general, discontinuity the more special, condition (Whittaker, 

 1956). As discussed by Bray (1956b), these studies and some other current 

 work (Goodall, 1954; Poore, 1955; Hale, 1955; Culberson, 1955; Bray, 

 1956a) are concerned less with associations than with ways of dealing effec- 

 tively with vegetational continua. 



The association thus ceases to be the clearly defined natural unit assumed 

 by Clements. Two further difficulties are suggested in relation to Clements' 

 associations. The first of these involves the dominance concept. The dominant 

 trees of a forest strongly affect conditions of life for other species of the 

 forest, but different tree species may affect the forest environment similarly. 

 It is not the species of the dominant but the total environment in the forest 

 which determines whether subordinate plant species can or cannot occur there. 

 The principle of species individuality applies to relations of subordinate 

 species to dominant species and total environmental complexes influenced by 

 dominants. Dominants do not control the distribution of other species and 

 organize them into well-defined groups of associates in the manner, or to the 

 extent, assumed by Clements (Whittaker, 1956). Furthermore, as has been 

 observed by European critics (Lippmaa, 1931 ; Braun-Blanquet, 1951, p. 560; 

 Ellenberg, 1954), the dominant species are often very wide-ranging ones, oc- 

 curring in most varied environments with most varied associates; the over- 

 emphasis of dominants by Clements and others is sometimes a rather crude 

 and superficial approach to relations of whole communities to environments. 

 Dominance cannot provide the simple, comfortable solution to many problems, 

 the master key for the recognition and classification of communities, indication 

 of community environments, and understanding of community function, as- 

 sumed by Clements. 



The other difficulty is in the treatment of associations as regional units. It 

 is easily shown that geographically, as locally, no two species of an association 

 have the same distribution (e.g., Billings, 1949). In consequence, the associa- 

 tion as a particular combination of species is a localized phenomenon (Lipp- 

 maa, 1933; Bourne, 1934; Cain, 1947; Knapp, 1948; Ellenberg, 1954); and 

 associations may be geographically, as locally, continuous with one another 

 (Walter and Walter, 1953). Species are distributed in relation to total environ- 

 mental complexes, not to climate alone; and species do not fit into sharply 

 defined groups corresponding to climatic regions and regional associations. 

 Approaches to generalization about geographic relations of species in com- 

 munities are being developed in Europe (Bocher, 1938; Meusel, 1939, 1943; 

 Walter, 1954b), but these provide no simple answers to the definition of asso- 

 ciations and determination of their boundaries. The major concepts of such 

 geographic treatment — species groups based on distributional relations, nat- 



