On the Bain and Drainage - Waters at Rothamsted. 
331 
The common salt applied contained 0'23313 gram of chlorine ; 
it will be seen that practically the whole of this was recovered 
in the drainage-waters. 
The chief cause of difference between this and the former 
experiment is the far greater activity of diffusion in the present 
instance. The chlorides when applied in solution to the wet 
soil began to spread downwards, though no drainage was taking 
place, and at the end of a week had made such progress that 
four days of percolation sufficed to bring chlorides into the 
drainage-water. That this early appearance of chlorides was 
due to their previous downward diffusion is proved by the small 
application of water at the surface necessary to cause this appear- 
ance. Had the chlorides remained at the surface it would have 
required the application of 850 grams of water to cause their 
expulsion, this being the amount of water necessary to displace 
the water already held by the soil ; but in fact the chlorides 
began to appear when only 480 grams of water had been applied. 
Nor was there only downward diffusion ; upward diffusion was 
also active during the 11 days of percolation ; and consequently 
the layer of water richest in chlorides was followed by a con- 
siderable amount of drainage containing chlorides in gradually 
diminishing proportion. The nett result of the 18 days' diffu- 
sion was that it required 1320 grams of water to expel the 
chlorides from the soil, whereas the chlorides were expelled by 
only 1000 grams of water in the previous speedy percolation. 
The chlorides are now also distributed throughout 690 grams 
of drainage-water, whereas previously the whole was contained 
in 150 grams.* 
We see from these experiments that the expulsion of the 
diffusible salts from a soil is effected most readily when the 
percolation is rapid ; that consequently a heavy rainfall, occur- 
ring in a few days, is far more dangerous in this respect than 
the same rainfall spread over a month. We see also that but 
for the action of diffusion, and other causes tending in the same 
direction, the soluble salts contained in a soil would descend 
on the application of rain in a well-defined band, and be 
suddenly discharged in the drainage-water ; whereas in fact the 
diffusion always going on in a moist soil tends to distribute 
the chlorides and nitrates equally throughout the mass of soil, 
and thus produces a considerable uniformity in the composition 
of the drainage-water. Evidence of the existence of bands of 
saline solution in the soil will, however, be found when we 
* The tardy expulsion of the chlorides from a wet soil was probably determined 
iu part by other causes besides ditfusion, but the view given in the text will 
suflSce for our present purpose. 
