DEPARTMENT OF BOTANICAL RESEARCH. 83 



great periods of glaciation, such as the Permian and the Pleistocene. 

 Instead of the evolution of a new flora and new cUmaxes, each advance 

 and retreat of the ice produced a corresponding shifting of the climax 

 zones in front of it. This shifting of the zones by which each was first 

 replaced by successively earlier or lower preclimaxes, and these in 

 turn by successively later or higher postclimaxes, constitutes a clisere, 

 i.e., a, successional development from one climax to another. The clisere 

 differs from the eosere in that no new flora is evolved, but existing 

 climaxes merely shifted, and differs from the sere in being a succession of 

 chmaxes instead of developmental stages which terminate in a relatively 

 permanent climax. The consistent application of developmental 

 principles has confu^med the original assumption that all features of 

 vegetation are the structural results of developmental processes, and 

 hence furnish direct evidence of the operation of the latter. This has 

 long been known to be the case in hydroseres, where the successional 

 movement from open water to the chmax is relatively rapid and sym- 

 metrical. However, these have been thought to constitute exceptions 

 and not to furnish the rule. The careful scrutiny of climax formations 

 throughout western North America during the last three summers has 

 shown the principle to be of universal application, in time as well as in 

 space. Even in the most static community, not only have the relations 

 of the various dominants and subdominants been found to record the 

 past development and to suggest that of the future, but it has also 

 proved possible to recognize developmental areas throughout, minute 

 and fragmentary as they often are. This is especially true in the chmax 

 zones of mountains, where surface and soil are extremely diverse and 

 where great differences occur in the smallest contiguous areas. This 

 successional analysis of dominance has made it clear that no two domi- 

 nant species are exactly alike in their demands, while it has justified 

 their grouping into communities on the basis of the similarity of their 

 responses. From this has come the far-reaching conclusion that the 

 dominant furnishes the key to the developmental study of vegetation, 

 as well as the chief objective of the experimental attack which must 

 accompany the latter. This has been definitized in the concepts of the 

 consocies and the consociation, which are respectively developmental 

 and chmax communities controlled by a single dominant. Consocies 

 fall naturally into developmental units, or associes, and consociations 

 nto climax units or associations, both of which are due to similarity of 

 physiological response. Thus, the recognition of the dominant as the 

 basis of investigation has made it possible to utilize both autecologic 

 and synecologic evidences in the study of vegetation, and to harmonize 

 them both in terms of development. 



