2G0 The Atmospliere as a source of Nitrogen to Plants. 
ammonia in rain falling at Paris, which coincided very nearly 
with those of M. Barral. The ammonia in rain-water he found 
to be nearly 30 times as much as that of the Seine. He also 
made two determinations of ammonia in the water of the sea,* 
taken at Dieppe, and, although the quantity present is very small, 
the fact of its existence is of the utmost interest and importance : 
since, considering the immense bulk of the ocean, and that it 
forms three-fourths of the surface of the globe, it may well have 
a most material effect in regulating the proportion of this consti- 
tuent in the air. The presence of ammonia in sea-water could 
not, however, surprise us when we reflect upon the amount of 
animal life existing in it, and the processes of animal decay 
which are continually going on there. M. Boussingault con- 
cludes this interesting Memoir by describing determinations of 
ammonia ho had made in snow. It was found, upon comparing 
water obtained by melting two portions of snow, one taken im- 
mediately it fell upon a stone terrace, and the other (from the 
same fall) after it had lain for 36 hours upon the soil of a con- 
tiguous garden, that the second contained ten times as much 
ammonia as the other. 
It is well known that snow has a most beneficial effect upon 
soils, and, amongst other causes, Boussingault believes that it may 
act in preventing ammoniacal emanations from the soil. He asserts 
his conviction that the excess of ammonia in the snow resting on 
the earth must liave been derived from the soil. His experi- 
ments, to have been conclusive, should have been made on quan- 
tities that had lain for an equal period of time ; the one in 
contact with the earth, the other not so : for from the porous 
nature of snow it may possess a power of gaining ammonia from 
the atmosphere which might by possibility lead to the observed 
differences. 
In the autumn of 1853, M. Boussingault made a number 
of determinations of ammonia present in rain-water, falling at 
Liebfrauenberg in Alsace, with the view of ascertaining how far 
previous experiments of M. Barral and himself, made in Paris, 
expressed the state of the atmosphere in the country and were 
applicable to agriculture. 
He observed, in the first place, that the ammonia present in 
the first rain that falls is much greater than in that wliich is 
subsequently examined. This result is, as was before explained, 
only what might be anticipated. In one case, after the rain had 
continued for 24 hours, the quantity of ammonia present in the 
water was reduced to almost nothing. From the same cause 
* These determinations of ammonia will be found in a note at the end of this 
paper. 
