308 GEORGE E. NICHOLS 



i.e., subdivisions of the association ''characterized by a single 

 dominant" (Clements 4, p. 129). Where the variation concerns 

 species of secondary importance it is similarly possible to dis- 

 tinguish different societies. Thus, a deciduous forest may be 

 essentially uniform in its structure throughout, so far as the 

 dominant species are concerned; or certain species may dominate 

 locally, gi'V'ing rise to more or less distinct consociations. Simi- 

 larly, the herbaceous or shrubby vegetation in such a forest 

 may, and usually does, vary locally, giving rise to more or less 

 distinct societies. In the case of both the consociation and 

 the society, the floristic composition of the vegetation affords 

 the sole criterion, although there may also be slight but rela- 

 tively inconsequential variations in the habitat. The con- 

 sociation and the society, then, represent floristic and not 

 ecological units. 



The association-type. The nature of any given habitat is 

 determined by a number of more or less definite, though not 

 always tangible factors. Wherever, within a given chmatic 

 region, a given set of habitat factors is duplicated, the same 

 type of habitat is the result. To express it algebraically: if 

 it is assumed that the nature of a given habitat. Hi, is determined 

 by the combined influence of the factors, Ai, Bi, Ci, and Di, and 

 if this fact be expressed by the equation Hi = f{Ai, Bi, Ci, Di), 

 then wherever the combination of factors Ai, Bi, Ci, Di recurs, 

 the value of H will be the same. Of course, the exact application 

 in practice of this criterion of the habitat would necessitate 

 a vastly more accurate knowledge of the factors which determine 

 its nature than is actually available, so that as a matter of fact 

 the parallels must be based to a large extent on superficial ob- 

 servations. But, in a general way, numerous parallel series of 

 habitats may be distinguished in every region. In comparing two 

 neighboring ponds, for example, the habitat occupied by the 

 Nymphaea association in the one may duplicate the habitat occu- 

 pied by the Nymphaea association in the other; or the habitat of 

 the Scirpus association in the one may duplicate the habitat of 

 the Typha association in the other; etc. Similarly, the individ- 

 ual habitats occupied in different Connecticut salt marshes by 



