LANDSCAPE 

 ARCH. 



PREFACE 



The present book is an endeavor to present the materials of the Rocky 

 Mountain flora in preliminary form from the standpoint of the experimental 

 ecologist. The latter is concerned primarily with the relationships of 

 "species" and their subdivisions as an organic expression or measure of 

 habitat differences, and of the competitive relations of the various forma- 

 tions. Whatever the taxonomic value of the numerous segregates of the 

 last decade or two, the fact that the binomial form conceals the relationship 

 to the original species, and that the segregate itself is based not at all or 

 only slightly upon habitat relations, makes them of little value to the ecolo- 

 gist. This condition is emphasized by the extreme difficulty of their field 

 determination and recognition. No attempt has been made to pass upon 

 the merits of segregates as such, but similarity and relationship have 

 been taken as determining the units used, with the conviction that the dif- 

 ferences will appear all the more clearly when habitat and formation have 

 been thoroughly studied experimentally. To the ecologist, ifr seems certain 

 that such experimental analysis of the unit must carry with it the regular 

 use of the trinomial, leaving binomials only for the unit as a whole, whether 

 capable of analysis or not. 



In spite of some quantitative study of the origin of new forms by 

 adaptation to the habitat, and some statistical study of variation from 

 habitat to habitat, during the past decade, the authors recognize clearly 

 the tentative nature of the units employed. While the latter agree in the 

 main with the "species" of Linne, and of Gray and the earlier American 

 botanists, the initial test of continuous variation or discontinuous adapta- 

 tion has merged a considerable number of these, and must be expected to 

 unite still more. The questions of a species, its inherited constancy, etc., 

 have not been raised, as this seems futile without continued experiment. The 

 units employed may be "species" or not, but at present they mean nothing 

 more than that the individuals or groups of individuals in a unit are more 

 nearly related to each other than to any other group. In fact, whenever 

 the curve of variation is continuous, it is felt that a unit is indicated, re- 

 gardless of the height of the modes. 



iii 



927 



