10 



ble, and their separation becomes to a great extent a matter of personal 

 judgment, and hence arbitrary. Nearly all of the so-called specific 

 characters are variable. The most common variations are in the 

 amount and distribution of the pubescence, the width of the glumes 

 and leaves, and the length of the awns. In most instances the amount 

 and distribution of the pubescence of the flowering glume, taken in 

 connection with other corelated characters of the plant, have been 

 found to furnish a fair basis for segregation, especially where these 

 characters are found to coincide with geographic distribution and 

 environment. It has been found, however, that a character which 

 seems fairly stable in one species may be much more variable in 

 another, so that no character in general can be said to have specific 

 value, but the value of each character must be determined by careful 

 study and observation of the plants throughout a considerable terri- 

 tory. Species founded on single characters are always artificial, rather 

 than natural. There are scarcely any of our species that do not show 

 intermediate forms. In some instances, of course, these are much 

 moi-e numerous than in others. The subgenus CeratocJdoa presents 

 the most numerous and perplexing forms. These seem to be largeh^ 

 due to the wide distribution and varied climatic and physiographic 

 conditions under which they are found. 



The advisability of the adoption of so many varieties may perhaps 

 be questioned by some as tending to make nomenclature too cumber- 

 some. As the chief value of the work, however, seems to us to lie in 

 bringing out as clearl}' as possible the degrees of relationship and rela- 

 tive importance of the various groups as well as their connection with 

 physical environment and geographical distribution, this course has 

 been followed as best serving the end in view. 



An attempt is here made to illustrate diagrammatically what appears 

 to us the relationships of the subgenera and species. 



The first group, Bromus proper, is made up entirely of forms ^vhich 

 have been introduced from Europe. 



brizseformis ' squarrosus 



ARVENSIS ■ JMtuluS 



secalinus racemosm 



I 

 scopanus hordeacem 



The second group, subgenus Neohrojnm^ represented by one species 

 and its varieties, shows no close relation with our other species, but 

 seems to approach nearest B. scopariiis and B. tectorum. 



The subgenus Steiwhroinus also consists entirely of introduced species. 



maximus 



I 

 madritens-is ruhens 



I 

 sterilis tedorum 



