NITRATES IN DRAINAGE WATEES 231 



from month to month in both the concentration of the drainage 

 water and the loss of nitrate. The concentration is at its 

 lowest in February, a time of year when the soil is still at a 

 very low temperature and has been thoroughly washed by 

 the winter's rains. In April the concentration has increased 

 but little, and this combined with a smaller percolation 

 results in the minimum loss of nitrates for the year in 

 this month. The rise in concentration is still slow until 

 July, when there is a considerable jump, the concentration 

 reaching its maximum in September. The maximum loss 

 of nitrates comes as soon after this point as the rain- 

 fall is abundant enough to wash through the soil; on the 

 average the greatest annual loss of nitrates takes place in 

 November, from which time onwards both concentration 

 and total loss diminish. All these results refer to soil 

 which is kept bare and uncropped, where in consequence 

 the percolation is at a maximum, and where also there 

 are no growing crops to take up the nitrates as they are 

 produced. The effects to be seen under more ordinary condi- 

 tions can be followed in the analyses of the drainage waters 

 from the Broadbalk wheat field, under each plot of which runs 

 a tile drain at a depth of from 2 feet to 2 feet 6 inches. These 

 drains all debouch into a cross trench at the bottom of the field, 

 and a record is kept of the occasions upon which there is any 

 flow from the drains. Determinations of the nitrates contained 

 in the waters are also made, but as it is impossible to with- 

 draw an aliquot sample of the whole flow of the drain, and as 

 also during any particular running the concentration is always 

 changing, the earlier discharge being sometimes stronger and 

 sometimes weaker than the later, it is impossible to make any 

 exact account of the quantity of nitrogen removed by each 

 drain. By combining, however, the results obtained over the 

 whole period of the determinations, an approximate idea of the 

 concentration of the discharge may be obtained. Other con- 

 siderations lead one to suppose that the discharge from the 

 drains does not represent, either in quantity or concentration, 



