torr] COOPER—LAYERING AMONG CONIFERS 377 
dynamics. For example, in the climax forest of Isle Royale there 
1S an appearance of thick reproduction, with a great preponderance 
of balsam in the young growth. Upon superficial examination 
one would conclude that reproduction by seed is taking place at a 
tremendous rate. Careful investigation reveals that a large pro- 
Portion of the apparent seedlings are in reality merely layered 
branches, some of them having originated from mature trees, and 
Many others being groups of connected shoots which have started 
from a single true seedling. The same situation was found by 
Clements (3) in Colorado. It is evident that the effectiveness of 
this method of reproduction will have an important bearing upon 
the course of the succession in the forest. The habit is of special 
importance in the region of timberline, where, according to authors 
quoted above, it is sometimes almost the only method of repro- 
duction. 
V. Physiological bearing 
There are also physiological problems involved in the phe- 
nomena of layering in this group of plants, which cannot at present 
be Satisfactorily settled. These problems relate to the theories 
of orthotropism and plagiotropism and their mutual relations. The 
Whole subject is at the theoretical stage, without adequate evidence 
mM support of any of the various hypotheses. In the process of 
layering, the rooting (when it occurs) is simple enough as a response 
to moisture and absence of light. The change from dorsiventral 
to radial symmetry is to be expected as a result of the tip becoming 
erect, being an adjustment to changed light relations. The change 
in direction of growth from horizontal to erect is the part that is 
difficult to explain. It is bound up with the agencies which cause 
lateral shoots, ordinarily plagiotropic, to become orthotropic when 
the terminal shoot is removed or damaged. In the process of layer- 
ing, it should be noted, this change takes place without antecedent 
removal of the main shoot. The case is thus somewhat different, 
but the same factors doubtless govern it. In the present state of 
knowledge relating to orthotropism, plagiotropism, and correlation, 
it will be useless to continue the discussion at length. One point 
however seems to be important enough in its bearing upon the 
