CORRELATION OF DISTRIBUTIONAL FEATURES. 515 



principal features of the climate for these two regions, brought together 

 for ready comparison. The widespread occurrence of evergreen forest, 

 extending into the northwestern, northeastern, and southeastern 

 corners of the United States, gives a very wide amplitude of conditions 

 for this type of forest. In 5 cases the amplitudes of conditions are as 

 great for the collective evergreen regions as for the entire United 

 States, and in 6 cases they are nearly as great. These wide amplitudes 

 are found among temperature and moisture conditions alike. The 

 narrowest amplitude is that of relative humidity, which ranges through 

 the upper half of its scale for the entire country. The importance 

 which humidity is here indicated to have for evergreen forest is borne 

 out by the detailed humidity extremes for each of the evergreen areas 

 (see fig. 25). 



The amplitude of the conditions found in the Deciduous Forest 

 region is narrower in every case than that of the collective evergreen 

 regions, and for none of the conditions does it approach the entire 

 amplitude for the country. It is also to be noted that the extremes for 

 the Deciduous Forest are, in nearly every case, well within the extremes 

 for the evergreen regions. In the number of cold days in the frostless 

 season the two have the same minimum, and in the number of days in 

 the longest normal dry period of the frostless season the two have 

 nearly the same minimum. The only case in which the two maxima 

 approach each other is that of humidity. 



If we disregard the diversities in the evergreen forests which have 

 led us to their separate treatment, it is possible to say that this type of 

 forest in general is capable of withstanding a much wider range of 

 climatic conditions than is the deciduous type. 



2. DISCUSSION OF THE OBSERVATIONS. 



We have now reviewed the observed correlations between the 

 general vegetation areas and some of the leading climatic conditions, 

 both from the point of view of the vegetation and from that of the 

 conditions. In the comparison of the amplitudes and extremes of 

 each single condition, as shown for each of the vegetational areas, it 

 has been possible to see to what extent that condition is unlike in the 

 several vegetations. It is obvious that a condition which ranges 

 through nearly the same values in two vegetations can not be looked 

 upon as one that is important in determining the optimum conditions 

 for each of these vegetations, nor as one that plays an important role 

 in controlling the limit between the two. It is evident, for example, 

 that the physiological summation of temperatures can not be held to 

 play a primary part in establishing the optimum conditions for Grass- 

 land, Grassland Deciduous-Forest Transition, and Deciduous Forest 

 (see fig. 22) . If, on the other hand, a given condition exhibits a sliding 

 scale of values for several adjacent vegetations, it is evident that such 



