1918] Waynick: A Statistical Study of Nitrification in Soil 253 



Unless otherwise stated, the probable error of any value has been 

 used directly as found in the discussion which follows so that there is 

 simply an even chance that the results so treated fall within, or 

 without, the limits of their respective probable errors. The effect of 

 multiplying the probable error by two, three, or any higher number, 

 may be found by referring to the reference cited above. 



Discussion of Experimental Results 



To express the results in concise form, table 5 has been inserted to 

 summarize briefly the data of the preceding tables. In the first column 

 the means of the various series are given with their respective probable 

 errors. In the case of residual nitrate. 2.7 milligrams in one hundred 

 grams of soil represent .0027 per cent of the soil or approximately 

 54 pounds of nitrate nitrogen per acre, considering only the upper six 

 inches of the soil mass. The probable error of the determination 

 amounts to .00005 per cent of the soil so that it is an even chance that 

 the mean of the eighty-one determinations is correct within one pound 

 per acre. In the subsoil, the mean of 14 pounds per acre, is within 

 0.6 pounds of the correct figure, on the same basis. The extreme range 

 amounts to 62 pounds in the surface soil and 18 pounds per acre in the 

 subsoil, both figures being of greater magnitude than the means of the 

 two series. It is not possible to translate the other figures into a 

 practical pounds per acre basis, since they represent laboratory treat- 

 ments. 



