RESEARCH METHODS IN STUDY OF FOREST ENVIRONMENT, 



2. The group averages for the loamy spruce soils and the granil 

 gravels are essentially the same. It therefore seems entirely legiti- 

 mate to consider both groups together and, as shown m diagram 1. 

 to express the relation of wilting coefficient to capillarity l>\ a si raighl 

 line. The nineteen cases show an average variation of 0.0333 from 

 the mean ratio of 0.202, or 16.5 per cent variation. 



3. Both from its mean value and from the facl thai the graph 

 which expresses this relationship passes through the main axis of 

 the system of coordinates it is evident that the capillary moisture 



it has been measured by the method described above is an entirely 



different expression from that used by Hilgard and by Briggs and 



Shantz. 



4. The relatively high wilting coefficients of the loamy soils having 

 the largest humus contents are believed to result from experimental 

 errors, largelv unavoidable, and due to the lack of capillar) con- 

 ductivity in soils which are particularly loose. This Lack permit 

 seedling to succumb in one region of the soil, while there maj be 

 considerable free moisture elsewhere. The two gravelly soils 

 show similarly high wilting coefficients also have high mois 

 equivalents, and it is thought from this that they were prob 

 richer than usual in permeable feldspar, which could not bold mucn 

 water but would probably hold it very firmly. 



