EVOLUTION OF ECOLOGICAL CONCEPTS 343 



liberal assumption a coherent, and apparently inclusive, deductive system of 

 vegetation interpretation. 



The structure of the system began to appear early in the century (Clements, 

 1905), was developed in its major features in the monograph on plant suc- 

 cession (Clements, 1916, 1928) and textbook of plant ecology (Weaver and 

 Clements, 1929, 1938), stated in definitive form in an essay on the climax 

 (Clements, 1936, 1949), and soon thereafter applied to the study of natural 

 communities through the plant-animal formation, or biome (Clements and 

 Shelford, 1939). In its course of about half a century it has accumulated in- 

 creasing criticism; although the most influential product of American 

 ecology, its influence has declined in the last decades. In considering the 

 meaning of this decline, the author has no desire to minimize the great con- 

 tribution which Clements made or to criticize Clements or his system for 

 the uncritical manner in which it was applied by some of his followers. The 

 object is not to show that Clements was mistaken, or that time and experience 

 have been unkind to his system, as they have been to others which were 

 brilliant in their times. It is to show, with this system as background, how 

 the climate of interpretation in a field of science has changed and seems to 

 be changing today. 



Changing views of ecological concepts. One of Clements' central con- 

 cepts is that of formation. Vegetation of the eastern United States seemed to 

 support his view of formations as regional and climatic units, each distinct 

 from others and unified within itself by a single dominant growth form. 

 Three great formations are easily recognized in temperate eastern North 

 America — the eastern deciduous broad-leaf forest, the grassland or prairie, 

 and the northern evergreen needle-leaf forest, or taiga. These seem indeed 

 distinct, natural, climatically determined units; there is only the problem of 

 the mixed broad-leaf-needle-leaf forests of the Lake States, to be treated as 

 either a transition between the first two or an additional formation. Experi- 

 ences elsewhere in the world have not supported the view that formations are 

 determined only by climate, that climatic regions and vegetation formations 

 must always be identical. The tropical grasslands, or savannas, surely among 

 the world's great plant formations, are thought to owe their existence to fac- 

 tors of both climate and soil (Beard, 1953). The formation is not simply a 

 unit reflecting climate alone, but a grouping of communities which are of 

 similar physiognomy and express a similar set of ecological conditions, of 

 which climate is only one (Beadle and Costin, 1952). It seemed simple 

 enough to recognize three major formations of temperate eastern North 

 America by one character only — dominance of a single growth form. But in 

 a broader view, many formations must be defined by mixtures of growth 

 forms, and many others must be distinguished by differences of environments, 

 not of growth-form dominance. Clements and others have been impressed 

 by the discontinuity between grassland and forest in eastern North America; 



