On the Rain and Drainage - Waters at Rothamsted. 35 
XLV.) we find the same increased proportion of chlorine in the 
waters from Plots 10-14,* and the increase is in much the same 
proportion, an addition of one-fourth to the chlorine of Plot 7, 
giving a figure between those actually found for Plots 13 and 
14. In the same Table we find that the lime and magnesia, and 
the total solid matter in the drainage-water participate in the 
increase. It is evident, therefore, that the drainage-waters at this 
side of the field are for some reason rather more concentrated 
than those collected at the other side. A reference to the plan of 
Broadbalk field (opposite first page) will show that the drain- 
pipes collect a portion of their water from an unmanured margin 
at the top and bottom of the field, and that the length of this 
margin is much greater at the right side than at the left side of 
the field, we should expect therefore that the pipes lying towards 
the left would deliver a rather stronger water than those lying 
towards the right. This, however, is not a sufficient cause for the 
difference observed, and we must assume, what is very likely, 
that a different proportion of surface water gains access to the 
pipes in different parts of the field. We have dwelt on this 
point somewhat fully, as it is important we should not attribute 
to the character of the manure an effect that may be merely due 
to the condition of the drain-pipes. 
On Plot 15 the chlorides are applied in October instead of 
in spring, as on the other plots ; this fact probably accounts for 
the increase of chlorine over that given by the adjoining Plots 
13 and 14. The quantity of chlorine applied to all these plots 
has been the same, but it has been subjected to different seasons of 
drainage. The water from Plot 15 was indeed much poorer in 
chlorides at the end of the three years tnan at the commencement ; 
while with Plots 13 and 14, and others receiving ammonium- 
salts in the spring, the contrary Avas the case. The two series of 
plots do not therefore admit of an exact comparison, and it will 
require a longer period of experiment to decide whether a larger 
amount of chlorine is generally contained in the drainage-water 
of the autumn-sown Plot 15 than in that of the spring-sown 
Plots 13 and 14 ; but from the results of Voelcker and Frankland 
we should expect that this will prove to be the case. 
We turn now to a part of the subject of far greater practical 
importance, namely, the evidence which the drainage-waters of 
Broadbalk afford, both as to the production of nitrates in 
the soil, and their removal from it by drainage. We must 
bear in mind throughout this part of the discussion that the 
* Plot 15 had at this time a different manure supplying leas chlorine. 
D 2 
