Sept, i, 1921 
Rate of Culm Formation in Bromus inermis 
811 
nation, while perhaps attractive, is rather too speculative for experi¬ 
mental proof. 
The problem at hand is simply the explanation of the increasing 
difference in number of culms per plant for the various groups with 
different initial culmage, assuming, from lack of knowledge, that initial 
differences were purely accidental (enviromental) rather than geno¬ 
typical. 
If an explanation is at hand for the increasing diversity, starting from 
initial differences in the seedlings (environmental in character), there 
Fig. a.— Plotting of average number of culms per plant at different dates of counting, arranged in four 
groups with regard to initial culmage. 
would scarcely be room for additional explanations based on genotypical 
diversity. 
In formulating another explanation, one may start with a general 
idea of the mechanics, let us say, of culm formation. For example, if 
a single culm is present, sooner or later an additional culm will be added, 
thus doubling the number. Later on the plant thus constituted will 
receive added increments, but the rate of increase of these future incre¬ 
ments will to a considerable extent have to be assumed. However, it 
is certain that each culm, new and old, can not go on adding one more 
culm to itself indefinitely at each stage of culm accretion. Such a rate 
would soon be self-inhibiting from purely spatial considerations. Very 
