THEORIES CONCERNING SOIL FERTILITY 



331 



tivation in Kansas during the fifty years, as briefly indicated by 

 the following tenth-year records: 



ACREAGE OF CEREALS IN KANSAS 



When we consider that eastern Kansas, the part first settled, 

 is in the humid section of the United States, and that the later 

 years include the records from the central and western parts of 

 the state where s'emiarid conditions prevail, it will be seen that the 

 average yields computed by Professor Spillman may serve best to 

 illustrate the possibility of drawing erroneous conclusions from the 

 use of general statistics unless full consideration is given to all 

 important factors. The explanation for the slight increase in the 

 average yields of the last nine years of the period, as compared 

 with the preceding ten years, is very possibly to be found in the 

 increased rainfall in the semiarid region, as is well illustrated by 

 the very interesting and very instructive diagram (shown on an- 

 other page) of the rainfall record at North Platte, Nebraska, for 

 the thirty-four years, 1875 to 1908 (Nebraska Bulletin 109, April, 

 1909), from which it will be seen that the ten years, 1890 to 1899, 

 included eight years below normal and averaged only 15.35 inches, 

 while the following nine years, 1900 to 1908, show but three years 

 below normal, and average 21.21 inches. 



It should be kept in mind that meat and dairy products bring 

 much larger returns in Maryland than in Kansas, and until the 

 well-situated, well-drained, and well-watered farm lands of Mary- 

 land and Virginia have been rehabilitated by these methods of 

 live-stock farming (which farmers have been familiar with for 

 centuries) ; until such soils as the Leonardtown loam, comprising 

 41 per cent of St. Mary County, Maryland, where, to quote the 



