BULLETIN OF TIE BUSSEY INSTITUTION. b15 
figures would indicate, for it was found that when water taken 
from the cistern was left to stand in open vessels in the green- 
house it lost some of its ammonia. Thus a sample left three 
weeks in a tall open glass jar (from 25th April to 13th May) con- 
tained 0.0005 grm. NH, to the litre at the beginning and 0.0003 
grm. at the close of the trial. Another sample left. standing in 
an open jar from 21st May to 7th June exhibited a precisely 
similar diminution, while a companion sample which stood in a 
closed bottle during this interval contained 0.0005 grm. NH, at 
the close as well as at the beginning of the trial. A half litre 
sample left standing in a wide-mouthed jar from 18th June to 
28th July showed only a trace of ammonia when tested with Ness- 
ler’s reagent, though the water had been found to contain 0.0005 
orm. NH, to the litre at first. A companion sample kept ina 
closed bottle beside the other during the same time was found to 
contain precisely as much ammonia at the end of the trial as at the 
beginning. Another sample of the cistern-water which had stood 
in an open vessel from 18th June to 6th August contained only 
0.00002 grm. NH, to the litre at the last-named date. A slight 
loss of ammonia from open vessels was invariably observed ex- 
cepting a single instance, where a sample of the cistern-water was 
left to stand in the greenhouse in a tall open jar for twenty-four 
hours on the 23rd April, 0.00048 grm. NH, to the litre being 
found both at the beginning and at the end; but when this sample 
of water had stood forty- -eight hours, the test showed only 0.00044 
grm. NH,. At the same ‘time a closed bottle of the water which 
had stood for forty-eight hours showed 0.0005 grm. NH, to the 
litre. Furthermore a sample of the cistern-water. kept in a closed 
bottle from 18th June to Sth Oct., 1875, was found to contain 
precisely as much ammonia at the end as at the beginning. For 
the sake of comparison, enough pure chloride of ammonium to 
amount to 0.0005 grm. to the litre was dissolved in pure distilled 
water and the solution was left to stand in an open wide-mouthed 
jar from the 16th August to 5th October, 1875, pure water being 
occasionally added to the jar to compensate partially for the loss 
by evaporation. When tested Oct. 5, the solution contained only 
0.00018 grm. NH, to the litre ; 7. e. the loss of ammonia amounted 
to 0.0003 grm. 
I was inclined to attribute this diminution of ammonia in the 
rain-water mainly to evaporation or perhaps to exhalation, in view 
of the constancy of the proportion of ammonia in the water of the 
cool underground cistern. Some of the ammonia might of course 
be changed to nitrites and nitrates; and this point was tested in 
the single experiment, where the water stood from 18th June to 
6th August, 1875. By means of the aluminum-soda method (as 
described by Wanklyn in his ‘‘ Water-Analysis,” edition of 1874) 
there was found in the sample of water which had stood in the 
open vessel as much nitrite and nitrate as would amount to 0.0002 
grm. NH, to the litre, while in the water which had stood in the 
