PENNSYLVANIA FIELD EXPERIMENTS 435 



These statements, if true, are of tremendous significance to 

 American agriculture, for they refer to the oldest experiments of 

 the kind in the United States; furthermore, the phosphorus- 

 potassium plot is repeated four times in every series, so that the 

 average results are from 16 different plots of normal soil every year 

 for twenty-five years, and they must be considered highly trust- 

 worthy. The small amount of limestone contained in this Penn- 

 sylvania soil can very easily be supplied to any other soil by the 

 direct application of ground limestone. 



It will doubtless be agreed by all that the results of the first few 

 years at the beginning of a rotation and fertilizer experiment are 

 not to be considered as comparable with the subsequent results. 

 There are several reasons for this; but, for the present purpose, 

 it is sufficient to consider that nitrogen may not have been a limit- 

 ing element for all crops at the beginning. The data given in 

 Table 8iP are not satisfactory for making any study of this special 

 point, because the averages for the unfertilized land include the 

 results from plot 8 which is represented to have received annual 

 applications of manure during the ten years previous to 1882, 

 because of which the addition of nitrogen alone appears (from 

 Table 8iP) to have actually decreased the crop yields, which is 

 not the case if we accept the system of comparison adopted for 

 Tables 78 to 81. 



It must be evident from every point of view that nitrogen was 

 not the limiting element for all crops at the beginning of these 

 experiments. It is evident, however, that phosphorus was the 

 principal limiting element at the beginning. 



Now, for the sake of simplicity, let us assume that from a given 

 type of very uniform soil (see Table 87) sufficient phosphorus will 

 become available during the season (1903) to meet the needs of a 

 54-bushel crop of corn (plot 102), while sufficient nitrogen will be 

 liberated for a 62-bushel crop. The application of nitrogen with- 

 out phosphorus could not be expected to appreciably increase the 

 yield (plot 103), while the addition of sufficient phosphorus with- 

 out nitrogen should increase the yield from 54 to 62 bushels, but 

 unless nitrogen was also supplied, the yield could not be expected 

 to go above 62 bushels. However, by applying nitrogen in addition 

 to phosphorus, the yield might be still further increased (as to 69 



