WILTING COEFFICIENTS OF DIFFERENT SOILS. 25 



WILTING COEFFICIENTS FOR DIFFERENT PLANTS GROWN IN A 



SERIES OF SOILS 



It is evident that in determining the wilting coefficients for different 

 plants great care must be used to provide uniform soil for the growth 

 of the plants, since a variation in the moisture retentivity of the soil 

 would lead to erroneous conclusions regarding the relative ability of 

 two plants to remove moisture. Furthermore, it is desirable to make 

 such comparative determinations in a number of soils of widely 

 differing texture and composition in order to eliminate a possible 

 influence of the soil used upon the final comparisons. Table XI gives 

 the wilting-coefhcient determinations for a large number of plants 

 grown in a series of soils, and embraces experimental work carried on 

 during a period of three years. Many of the determinations were 

 made before the necessity of some of the precautions which have been 

 mentioned in the first part of the paper were fully appreciated; conse- 

 quently there is a greater fluctuation in the results of the individual 

 determinations than otherwise would have occurred. 



The determinations have been carried on in 20 different soils, ranging 

 from a coarse sand to a clay loam, and the data as presented include 

 all determinations which have been made upon these soils. The deter- 

 minations for the different plants are grouped with respect to the soil 

 in which the determinations were made, the soils being arranged in 

 the order of increasing wilting coefficients. A direct comparison of 

 the wilting coefficient obtained for different plants with the same soil 

 can thus be readily made without reducing the observations in any 

 way. In using so many different plants the number of observations 

 made with each plant have necessarily been limited, and too much 

 emphasis should not be placed upon the differences shown in the mean 

 values of short series in which the number of observations has been 

 insufficient to establish the probable error. It should also be borne 

 in mind that the probable error indicates only the accidental errors. 

 Constant errors, due to any cause, such as an inadequate root dis- 

 tribution in certain species of plants, would appear as real differences 

 in the wilting-coefficient determinations for the different plants, and 

 such errors would not be indicated by the probable error. 



230 



