No. 468] AMERICAN FOREST CENTERS. 877 
In using the term * center of distribution " it is not implied 
that the plants have necessarily spread from these centers, but 
that the complex of climatic factors most favorable to the devel- 
opment of this type of vegetation is here localized, and that as 
we depart from such centers we find conditions more and more 
unfavorable. This implies the elimination of. such species as 
are most rigidly dependent upon definite conditions. Compare 
for example (Fig. 1) the distribution of Pinus strobus and Tsuga 
canadensis with that of Picea mariana or Larix laricina and 
note the intermediate dispersal of Adzes balsamea. | Fig. 2 shows 
the distribution of several other species and still further rein- 
forces the suggestion that the St. Lawrence basin is a definite 
center about which is distributed a unique type of forest. It is 
floristically related to the forests of other parts of the continent 
but ecologically and climatically it is distinct. 
The recognition and separation of these centers is of the 
greatest importance at the present time when there is so much 
activity along physiographic ecological lines. That there is a 
natural succession of plant societies in a given locality has long 
been recognized. Cowles (:o1a) has shown that this succession 
may be correlated with the physiographic development of the 
region, because soil structure, water content, and slope are 
largely determined by it. Already the applicability of this 
hypothesis has been demonstrated in many parts of the United 
States. 
But thus far little attention has been paid to the relation 
which the successive plant societies found in an area bear to 
their centers of distribution. In other words, the societies have 
been studied from their physiographic, but not from their geo- 
graphic aspect. Such a geographic outlook, however, not only 
aids in the selection of the characteristic plants of the local hab- 
itat, but also throws light on the relative importance of the sev- 
eral societies. The geographic point of view is also necessary 
to furnish a suitable basis for comparing local meteorological 
data. Unless the climatic conditions of the centers of distribu- 
tion of the societies are pointed out, the meaning of local climatic 
data is not apparent. ` Further, it is probable that in many areas 
the societies are not all members of the same geographic center, 
