44 
present in cistern waters, in so far as this may he estimated 
by the chemical processes at present in our possession. 
The method which I have adopted consists in measuring 
the amount of free and of albumenoid ammonia, and the 
amount of nitrogen existing as nitrates and nitrites : from 
these data we may deduce, at any rate comparative measure- 
ments, of the purity of various waters. 
In order to prove conclusively, for my own satisfaction, 
that if sewer gases be absorbed by water their presence will 
be indicated by a marked increase in the quantities of 
ammonia, free and albumenoid, obtained from that water 
on analysis, I carried out the following preliminary experi- 
ment : — 
A quantity of distilled water, free from ammonia, was 
placed in a porcelain basin, which was covered with porous 
paper, and suspended at a short distance above the liquid in 
a sewer which received the refuse from a very large area, 
chiefly occupied by dwelling-houses, in Glasgow. After 96 
hours, the free and albumenoid ammonia were estimated 
with the following results : — 
Free ammonia =0 '60 mgms. per litre = parts per million. 
Albumenoid „ —Ob 4 „ „ = „ „ 
It is thus evident that the absorption of sewer gases by 
water causes a marked increase in the quantities of ammonia 
obtained on analysis. 
The method adopted for the estimation of ammonia was 
the well known one of Wanklyn and Chapman: the method 
for the estimation of nitrates was that described by Thorpe 
in the Journal of the Chemical Society for June, 1873. 
This method consists in evaporating the water, along with 
a fragment of ignited quicklime, to a small bulk, and then 
evolving the nitrogen, as ammonia, by the action of zinq 
coated with a deposit of spongy copper, at a boiling heat. 
In selecting the waters for examination I endeavoured, as 
far as possible, to obtain typical samples : in this endeavour 
