April 1933] FERTILIZER EXPERIMENTS ON "RUN-OUT'* HAY LAND 9 



superphosphate or manure. This is true in spite of the fact that the 

 soils of the plots receiving limestone (L) did not reach a point much 

 above pH 6.1. 



It is rather interesting to note, however, that the first five cuttings 

 showed a slight decrease for lime and the last five a slight gain, al- 

 though these losses and gains are not significant. Future cropping 

 may possibly justify the use of more lime than was applied as a basic 

 treatment, but present farming conditions do not warrant an expendi- 

 ture which does not begin to pay dividends within a five-year period. 



As lime leaches out of a soil in drainage water more readily than 

 any other important substance purchased by farmers for soil improve- 

 ment, it is probably wise to prevent such loss by using minimum re- 

 quirements for the crop to be grown. This means two tons of lime- 

 stone for alfalfa and sweet clover for most New Hampshire soils, and 

 less than this amount in certain sections of the lower Connecticut valley. 



For crops other than alfalfa or sweet clover it appears probable that 

 an application of less than two tons of limestone will prove more eco- 

 nomic al. 



Figure 1 presents graphically the correlation between annual rain- 

 fall and yields of alfalfa hay in representative treatments. Rainfall 

 records were taken at Durham which is, by airline, eight or nine miles 

 from the field in question. They cover the six months' period from 

 April 1 to September 30 of each year. 



On this soil there appears to be a very close correlation between yield 

 and rainfall. 



A study of the comparative yields of the three series of plots in the 

 graph reveals, too, that the productiveness of the check plots fell off 

 considerably and the K or potash plots slightly, whereas the yields of 

 the heavily fertilized, limed and manured plots held up well. 



This is due, in the main, to the fact that the stand of alfalfa was 

 maintained more fully on those plots receiving potash and other ma- 

 terials over the basic treatment. 



In 1928 the crop w^as sampled by plots for the determination of nitro- 

 gen. The samples were taken from the hay in cocks, after curing. Du- 

 plicate samples from the same plot agreed fairly well in nitrogen con- 

 Lent, but the checks of moisture were, in many cases, far from satisfac- 

 tory. For this reason, no attempt has been made to calculate the total 

 yield of protein ; only the per cent, of protein in the dry sample is re- 

 ported. 



The difficulties in obtaining accurate figures for the total yield of any 

 constituent would make 'it seem advisable in future samplings to select 

 plants from various parts of the plot while the crop is still standing. 

 This would prevent accidental contamination of the sample with plants 

 other than the main crop, and should yield the most valuable data as 

 to the effects of the treatment on the composition of the particular 

 plant being studied. 



The average protein content of the samples from the plots under each 

 treatment and for each cutting are listed in Table IV. It is doubtful 

 if the differences in treatment had any effect on the protein content of 



