1 2 8 TILL A GEMO VEMENTS OF SOIL WA TER [chap. 



the soil. For this reason, in the Rothamsted experi- 

 ments, the plot growing wheat continuously has given a 

 greater crop per acre per annum than the plot fallowed 

 and sown with wheat in alternate years, though the 

 wheat crop following fallow has always been larger than 

 the crop grown the same year on the unmanured plot. 



Of course the average yield on the fallowed ground 

 was twice the above figures, i.e.> \j\ bushels of grain and 

 1595 lbs. of straw, but it was only grown every alternate 

 year. 



That the autumnal rainfall is the great factor in 

 determining whether a bare fallow shall be profitable 

 or not to the following crop, may be well seen by a 

 further examination of the results obtained at Rotham- 

 sted on these plots, by comparing the crops with the 

 percolation which took place in the autumn previous. 



The percolation through 60 inches of bare soil for 

 the four months, September to December inclusive, as 

 measured by the drain-gauge previously described on 

 p. 85, amounted on the average to 6-45 inches for the 

 31 seasons, 1S70-1901. If, then, we divide the harvest 

 years into two groups according as the autumnal 

 percolation is above or below the average, and allot 

 to each year the crops on the continuous wheat and 

 wheat after fallow plots for the harvest following the 

 given percolation, we shall obtain the following 

 average results, which show in group 1 the mean 

 crops following autumns of less than average percola- 



