14 BULLETIN 122, U. S. DEPARTMENT OF AGRICULTUEE. 



DISCUSSION OF THE RESULTS. 



The most important generalization to be drawn from Table I is 

 that most of the rarer elements tested for were found in all the soils. 

 They seem to be rather evenly distributed among the various soil 

 types and provinces. The barium of the York silt loam soil and subsoil 

 is remarkably high. There appears to be some quantitative relation 

 between barium and potassium. Strontium does not appear to be 

 associated with barium, nor do these elements appear to be com- 

 bined with sulphur. Molybdenum was proved to be present in two 

 cases only. 



Rubidium was detected in all soils examined. In addition to the 

 soils given in Table I it has been found in Colorado sand from Colorado, 

 Knox silt loam from Missouri, Oswego silt loam from Kansas, and 

 Greenville sandy loam from Georgia. Caesium was proved to be 

 present in only one soil, Colorado sand, from Greeley, Colo. Since 

 Vernadski * has found caesium to be widely distributed in feldspars 

 and micas it would seem that there must be small amounts present 

 in all soils, since all soils contain large quantities of either feldspars 

 or micas. 



The amounts of the rare earths precipitates were too small to 

 separate into simpler groups or elements. The precipitate probably 

 contained cerium for the most part, with a smaller amount of thorium. 

 Thorium has been reported in an Italian soil to the amount of 1 

 part of thorium oxide to 60,000 parts of soil by Blanc, 2 and Joly 3 

 has found it in sedimentary rocks in amounts averaging about 

 1.2 XlO -5 grams per gram of rock, or 1 part in 70,000. 



Nickel and cobalt are probably generally present in soils; the 

 figures given do not have quantitative significance, however, and 

 are subject to the doubts referred to in the paragraph on methods. 

 Further work is needed to establish the amounts of these elements 

 present in soils and also to prove definitely the presence of copper in 

 anything but the merest traces. 



Since the number of soils in the Coastal Plain and Limestone 

 Valley and Upland provinces analyzed was small, the partial analyses 

 of such soils reported by Failyer, Smith, and Wade 4 are given in 

 Table II for the purpose of comparison. 



Soils of the Glacial and Loessial province are higher in calcium, 

 magnesium, sodium, and organic matter than those of the Piedmont 

 Plateau or Coastal Plains. Whereas most of the soils of the Piedmont 

 Plateau are low in potash, the large amount in numbers 19, 20, 21, 

 and 22 brings the average up to that of the glaciated areas. 



iBul. Acad. St. Petersburg (1909), 821. 3 Phil. Mag., 20, 353. 



2 Atti. Accau. Lincei., 17, 1, 101. * Bui. 54, Bureau of Soils, U. S. Dept. Agriculture (1908). 



