On the Rain and Drainage - Waters at Rothamsted. 
53 
In default of more accurate knowledge of the amount of 
drainage from the land of the experimental wheat-field, the 
drainage through the GO-inch soil-drain-gauge is in the follow- 
ing, as in the preceding estimates, taken as the basis of calcula- 
tion. We have, however, the record of this only for the last eleven 
years of the thirty. For the preceding seventeen (as well as 
for the last eleven) years we have the record of the rainfall at 
Rothamsted, and for the first two years we adopt the amounts 
for a neighbouring station (Xash Mills) ; so that thus we have 
the rainfall for the first nineteen years. Then, each year being 
divided into the characteristic periods — from autumn sowing 
to spring sowing, from spring sowing to the end of May, from 
June 1 to harvest, and from harvest to autumn sowing again — 
the rainfall for each such period for each of the first nineteen 
years is taken ; and the drainage of each period is assumed to 
be the same as with the nearest corresponding rainfall for like 
periods during the eleven years for which the record of both 
rainfall and drainage is available. 
As to the composition of the drainage, the amounts of which 
are so estimated, we have for the twenty-five years during which 
the ammonium-salts were sown in the autumn, only the few 
determinations by Dr. Voelcker and Dr. Frankland, on samples 
collected in only a few of the twenty-five years. Upon these 
we have to rely in estimating the composition of the drainage 
of two of the most important of the four periods of the year 
into which each of the twenty-five years is divided ; and for 
the other periods, less influenced by the time of sowing the 
manure, or by growth, average figures are adopted from the 
more recent determinations. For the second year of the thirty, 
when the ammonium-salts were spring-sown, and for the first 
of the last four years of spring sowing, the composition of the 
drainage of the different periods of the season is calculated 
according to the average results for the corresponding periods 
of the last, or succeeding three years. For these three years 
themselves, an almost complete series of actual determinations 
is available. 
The following Table (Lll.) gives the so estimated losses of 
nitrogen by drainage, in lbs. per acre per annum, for each of the 
different series of vears. It also jjives the average for the whole 
thirty years ; and for comparison the mean for the two years 
from the time of spring sowing in 1879 to the same period in 
1881. In the last two columns are given, for the thirty years, 
and for the two years, the estimated losses from each plot 
receiving nitrogenous manure over Plot 5 with mineral manure 
alone. 
