PREFACE 



this material convinced me that it is futile to attempt to distin- 

 guish centres of variation, or to recognize such species as Ha- 

 benariajragrans, Platanthera huronensis, and P. graminea. 



The influence of environment is too frequently ignored by 

 those who pay special attention to shght deviations from the 

 condition represented by the type. Habenaria hyperborea and 

 H. dilatata, which were originally described from plants collected 

 in northern localities, are extremely variable in their vegetative 

 organs. From the characteristic dwarf state met with in the ex- 

 treme northern part of the range, to the slender or robust tall 

 state common southward, every gradation for a complete series 

 may be found. This fact, in conjunction with the information 

 offered by Gaston Bonnier, is very significant.^ Bonnier experi- 

 mented with single individuals which he divided into two. One 

 part of the original plant was grown on the Alps or Pyrenees, 

 the other on low land. In a short time the alpine half assumed 

 the familiar dwarf habit of alpine plants. For an example of the 

 difference between the two halves of a plant treated in this man- 

 ner, Helianthemum vulgare should be consulted in the illustra- 

 tions of Bonnier's paper. 



In the "Evolution of the Orchidaceae," in the Orchid Review 

 (February, 1910), R. A. Rolfe treats the Habenaria group under 

 two subtribes, namely, Gymnadeniege and Habenarieee, the for- 

 mer having one and the latter two distinct stigmas. He says that 

 Platanthera and Gymnadenia are sometimes united with Ha- 

 benaria, an arrangement which makes of Habenaria one great 

 chaotic aggregate whose characters cannot be defined with any 

 degree of precision. This statement leads to the conclusion that 

 the characters on which the three genera depend for purposes 

 of classification are not sufficiently clear to allow subgeneric 



^Ann. Sci. Nat., 7 Ser., t. 20. 



[ix] 



