﻿54 BOTANICAL GAZETTE [January 



societies. The ultimate cause of all this varied vegetation must 

 be something more particular, something which will affect the 

 individual plant. For such a region as this, this sometkmg 

 must exist in the nature of the soil ; climatic factors cannot 

 explain differences in such a small area; and the historical fac- 

 tor is broad and general, like the physiographic one, and hence 

 is not ultimate. That local differences in vegetation are due to 

 soil factors has been practically proved before this, and the proof 

 is strengthened by the present study. The physiographic 

 hypothesis explains how it is that various soils may be physi- 

 cally and chemically different. But, after this is explained, the 

 question with which we have to deal lies still untouched : What 

 is it in the nature of the soil which determines the distribution of our 

 plant societies ? 



Now, by *' nature of the soil" two things may be denoted, 

 and only two, i. e., the physical nature and the chemical nature. 

 But neither of these can influence the plant per se. Either one 

 of the features may be effective, however, through soil water. 

 Water is the only feature of the soil which comes in direct con- 

 nection with the vitality of the plant. The chemical nature of ■ 

 the soil may be directly effective through the nature of the dis- 

 solved substances which enter the plant, or indirectly, through 

 osmotic pressure. Its physical nature may be effective through 

 the retention or non-retention in the soil of the water itself. 



So far we may go a priori; beyond this, tests must be made. 

 The nature of the soil water from various soils in various posi- 

 tions must be carefully determined. From these determinations 

 will be shown how much truth or falsity there is In the explana- 

 tion here offered, that the nature of the soil water is not usually 

 a decisive factor for such a region as this. 



Also, by careful tests the ability of various soils to retain 

 water must be determined, and these determinations recorded 

 with the vegetation found growing where the tests were made. 

 Thus, and thus only, can the hypothesis here offered be tested, 

 /. e., the hypothesis that the decisive factor in plant distribution over 

 a small glaciated area is, in most cases, the moisture -retaining power 



