Chemical Changes in Mixtures of Sewage and Sea-Water. 341 
the unavoidable experimental errors, the result may be misleading. 
In all the determinations (with the exception of the nitrites), the 
nitrogen was eventually obtained as ammonia, and the latter 
estimated by Nesslerizing. 
In determining both the free and albuminoid ammonia, a modifi- 
cation of the usual process was employed. 500 c.c. of the mixture 
of sewage and sea-water was distilled until 200 c.c. had passed over 
into a graduated flask. The contents of the latter were then well 
shaken, and 50 c.c. Nesslerized for the free ammonia. In a similar 
manner, after adding the alkaline permanganate, a further 200 c.c. 
were distilled off, and 50 c.c. Nesslerized as before. It is clear that 
the results in each case had to be multiplied by 4 to give the free 
or albuminoid ammonia in the original 500, and by 8 to obtain 
parts per 100,000. 
It is obvious that a small error in the amount of ammonia 
found in the 50 c.c. of distillate becomes greatly magnified in the 
expression for the final result. Thus, an error of 3 c.c. of the weak 
ammonia solution used for the determination corresponds with -004 
parts per 100,000 or :0033 of nitrogen, and such an error is quite 
probable. 
In the case of the nitrates, the error becomes magnified five 
times, as 200 c.c. of the original fluid were treated with the zinc- 
copper couple, and the resulting ammonia determined in the whole 
of the distillate. 
In view of these facts, a more accurate—or, perhaps, we should 
say a less misleading—-method of expressing the results would 
be to give the actual number of cubic centimetres of the weak 
ammonia solution used in the different determinations, and this 
we have done in the following Table. 
(TABLE. 
