for Ttcenty Years in succession on the same Land. 343 
would represent a loss of nearly 5 (4"88) lbs. per acre for each 
inch of rain so passing. At this rate it would require little more 
than 7 (7'38) inches of rain to pass beyond the reach of the 
roots to account for the whole loss of nitrogen observed in the 
case of the wheat plot No. 7. 
We have said that the actual amount of drainage is unknown ; 
and since, in the case of the land in question, the subsoil of clay 
rests upon chalk at from 6 to 10 feet from the surface, and there 
is, therefore, natural drainage constantly going on, no gauging 
of the flow of the pipes, however exact, would indicate the total 
amount of water passing. Other experiments at Rothamsted. 
have, however, proved, that from one- third to one-half of the 
annual rain may pass below 40 inches. Supposing only one- 
third of the total fall so to pass, an average of from 8 to 9 inches 
of rain would annually drain away, by far the greater proportion 
of which would go off during the autumn and winter months. 
The quantity and composition of the drainage-water here sup- 
posed would obviously be sufficient to account for more than the 
whole of the loss of nitrogen from Plot 7 as above indicated. 
On the one hand, however, some allowance in the way of de- 
duction must be made for the amount of nitrogen as nitrates and 
nitrites in the drainage, due to accumulations within the soil prior 
to the period included within the estimate, or to other normal 
annual sources ; but whether, with the large annual supply of 
nitrogen by manure, and the much more active root development, 
in the case of Plot 7, the amount of nitrogen in the drainage- 
Avater from that plot, due to sources other than the annual direct 
supply of nitrogenous manure, would be as much as that indi- 
cated in the drainage from either plots 3, 4, or 5, may be a 
question. On the other hand, the proportion of the drainage to 
the rain-fall, in the case of the soil in question, would probably 
average more than one-third, which amount only is assumed in 
the above estimate. 
Although the selection of samples sent to Dr. Frankland was 
very fortunate, so far as the illustration of the wide difference in 
the composition of the drainage from the same plot at different 
times is concerned, his results are, on that account, the less 
directly available as a means of forming a judgment of the pro- 
bable average composition of the drainage throughout any par- 
ticular season of the year. To this end it would be desirable to 
have had results relating to the period between January 5 and 
May 18, 1872; and again to that between October 26, 1872, 
and January 19, 1873. Still, taking Dr. Frankland's results as 
they stand, the mean proportion of nitrogen as nitrates and 
nitrites in the samples of drainage from Plot 7, collected on 
January 5 and October 26, 1872, and on January 19 and 
