browsing by wildlife, trampling, reactions of plants, erosion, de- 

 position, repeated flooding, etc. For the direction of change to be 

 maintained as a retrogression, the influence of the causative agent 

 must continue at a certain minimum intensity, such as a uniform 

 rate of deposition, persistence of polygon formation, or mowing. 

 There may be an actual reversal in directional change, such as 

 the change from forest to prairie in lUinois,^^ but Clements'^ ^ 

 denies the existence of a directional change which repeats in re- 

 verse order the stages of a succession. A retrogression is not to be 

 confused with a downgrade phase of a cyclic replacement change 

 which is a normal developmental process. Other terms used for 

 retrogression are "degradation" and "destructive change." These 

 terms usually indicate that there are decreases in one or more of 

 the attributes of succession, that is, complexity, diversity, stability, 

 productivity, self-maintenance, uniformity within and between 

 stands, and soil maturity. 



Some directional changes may be considered as deflected de- 

 velopment.^^ For example, forest, shrubs, and savanna derived 

 from the tropical rain forest and subjected to grazing and recur- 

 rent fire, are the result of deflected succession and lead to 

 apparently stable climaxes.^^' The directional change that is 

 usually expected is modified to a significant degree. 



The upgrade and downgrade phases of a cyclic replacement 

 change and the intercommunity cyclic change are not to be con- 

 fused with directional changes. Differentiation must be based on 

 determining that a phasic cycle or an alternating cycle is opera- 

 tive rather than a continuing directional trend. The differentia- 

 tion between the latter and a phasic cycle is confounded to a de- 

 gree because a phasic cycle may be operative in succession, such 

 as the alternate establishment and disintegration of the Danthonia 

 spicata-Cladonia cristatella-C. subtenuis-moss community until it is 

 finally invaded by Andropogon virginicus or Pinus virgimana.^^^'^ 



In order to determine that a directional change, such as suc- 

 cession, is taking place, it is necessary either to follow the course 

 of change on a given site, such as permanent sample areas (see 

 Chart 4, p. 104), or to use circumstantial evidence, such as the 

 presence of invading species or of relict species of an earlier stage. 



Habitat Patterns, Changes, and Climax • 155 



