ORCHIDACE.E 



would appear of generic value would hardly appear so in another. 



Characters which are to serve for generic distinction are not 

 infrequently chosen with total disregard of specific forms through- 

 out a wide range, and this is most likely to take place in the 

 segregations made from amphigean genera which comprise nu- 

 merous polymorphic species. In a localized flora made up in great 

 part of vagrants which have become introduced into a country 

 from distant, geographically distinct regions, those systematists 

 who lack opportunity to study in large herbaria where general 

 and fairly complete representations are to be had of the flora of 

 adjacent territory, and who confine themselves to the plants of 

 special geographical regions, are inclined to establish genera on 

 characters which a broad knowledge of a given group would 

 clearly show were scarcely of generic weight. 



In southern Florida, for example, we find a remarkable or- 

 chid flora largely made up of West Indian immigrants. Here 

 the genus Epidendrum is represented by about ten species, be- 

 longing to at least six very distinct sections as follows : Epicla- 

 dium, Encyclium, Hormidium, Osmophytum, Amphiglottium, 

 and Euepidendrum. With their few representatives these sec- 

 tions might well be regarded as distinct genera,^ and would ap- 

 pear as six very natural groups if no other species outside of 

 Florida were known. Here too a noteworthy occurrence would 

 perhaps justify a still broader treatment, as the single species 

 representing in Florida the section Osmophytum has three well 

 developed anthers and is unknown in its normal state in this re- 

 gion. If the components of this triandrous race were recognized, 

 on the basis of the gynoecium, as constituting a distinct genus, 

 we would then have a species which had changed its generic or 

 even its tribal character, while its specific characters remained 



1 Hormidium is so recognized. 



[5] 



