On the Rain and Drainage - Waters at Rothamsted. 333 
reached their maximum in the seventh extract, and were all 
removed at the ninth. The nitrates thus appeared later than 
the chlorides, and ceased sooner ; the nitrates were spread over 
five extracts, while the chlorides occupied eight. Nitrites were 
found in the drainage-water. 
The whole of these facts find their explanation when we 
look at the quantity of nitrate recovered in the drainage. The 
nitrate of sodium employed contained 0'09iy8 gram of nitrogen ; 
of this only 0*01920 gram, or 20'9 per cent., was recovered in the 
drainage-water. It will be recollected that in the corresponding 
experiment with chloride of sodium practically the whole of the 
chlorine was recovered. How has this serious loss of nitric acid 
occurred ? Clearly by reduction of the nitrates in a water-logged 
soil, destitute of free oxygen. This reduction in question has 
been effected by the organic matter of the soil, and has resulted 
in the formation of carbonic acid gas. A part of the nitric acid 
has probably been reduced to ammonia, while a considerable part 
of the nitrogen has most likely taken the form of nitrogen gas. 
The soil employed was by no means rich in organic matter ; 
but tlie perfect consolidation of the soil, the removal of air by 
the pump when water was first poured on, and the fact that the 
soil was always afterwards covered with water, afforded oppor- 
tunity for the consumption of all available oxygen, and then for 
the reduction of the nitrates present. The experiment was con- 
ducted in April ; the temperature was therefore not high. 
To confirm these results, the soil already treated with chloride of 
sodium was made use of for a second experiment with nitrates. 
The quantity of nitrate of sodium used was double that pre- 
viously employed. Instead of waiting a week after the applica- 
tion of the nitrate, percolation was started a few hours after its 
addition to the soil ; less opportunity for reduction was thus 
afforded. The results appear in Table XXX., p. 334. 
The development of cracks in the soil, the retardation of 
drainage, and the production of nitrites were observed as before. 
The nitrate of sodium employed had contained 0"18396 gram of 
nitrogen; of this 0*10305 gram, or 56 per cent., was recovered in 
the drainage-water. The absolute loss was really, however, 
rather larger than in the previous experiment with half the 
quantity of nitrate ; the loss then was 0'07178 gram, while now 
it amounted to 0'08091 gram of nitrogen. 
The reduction of the nitrates in soil to ammonia and gaseous 
nitrogen, when oxygen has been excluded, has been observed by 
Schloesing * and others ; the fact is of considerable agricultural 
importance, as showing the loss of nitrates, and even of soil 
nitrogen, which may occur in ill-drained soils in wet weather. 
Convptes Bendus, Ixxvii. 353. 
