358 0. D. von Engeln — Mineral Plant Foods. 



present, an average from 15 samples. An average of nine 

 samples of brown silt loam subsoil of similar depth and location, 

 but from the youngest, or Late Wisconsin glaciation, shows 

 87*6 per cent of the phosphorus content to be acid soluble. 

 The difference in solubility of the phosphorus in these subsoils 

 is therefore 6*4 per cent. In the surface portions, 0-7", of 

 these same soils a more marked difference is apparent. Of the 

 Lower Illinoian surface soil 80 per cent, of the phosphorus is 

 acid soluble, in the Late Wisconsin 89 - 7 per cent, a difference 

 of almost 10 per cent. This greater difference in solubility is 

 probably influenced by the fact that the Late Wisconsin soil 

 has its higher, total phosphorus content in the surface layer; 

 the concentration of this constituent being much more nearly 

 of the same magnitude in the subsoils of the two glaciations. 

 The average acid solubility of the total phosphorus in ithe 

 widely different soils of Maryland is 56 - 5 per cent ; and only 

 three soil series (?) of the sixteen given show a percentage solu- 

 bility of over 80 per cent. This is in marked contrast to the 

 comparative uniformity shown by the extreme types of the 

 glacial soils. 



If these tentative comparisons and conclusions serve to point 

 out the paucity of reliable data on the chemical composition of 

 soils in the United States and the need of further systematic 

 investigation along such lines, before broad generalizations are 

 made regarding the effects of varying amounts of mineral 

 plant foods on soil fertility, their purpose has been, in the 

 main, achieved. They further indicate a relationship between 

 chemical composition and the geologic-physiographic conditions 

 of soil origin and state. On this basis soil classification should 

 follow physiographic principles, and soil sampling should be 

 done in the field by investigators trained in geology and physi- 

 ography. The names of soil series might well correlate with 

 recognized physiographic and geologic formations in the United 

 States. A large scale map of the United States and Southern 

 Canada showing the areal distribution of acidic and basic igne- 

 ous rocks, sandstones, shales, and limestones, irrespective of 

 position in the geologic time scale, would give a basis for the 

 correlation of soil surveys having regard for such a classifi- 

 cation. On the same map, or on different ones, should be 

 shown the areas of glaciation, generalized drainage features 

 (topography), and the climatic regions of the United States, to 

 serve as a further basis for such classification. There is no 

 attempt to argue here that soil fertility will closely correlate 

 with provinces so marked out. Soil texture, bacteriological 

 considerations, etc., may be the fertility determining factors, 

 without rendering invalid the rationality of such a classification 

 of soils. 

 Cornell University, June, 1911. 



