302 Scientific Proceedings, Royal Dublin Society. 
hence the terms ‘ carbon fermentation’ and ‘ nitrogen fermentation ’ 
which I have employed. When the ascertained consumption of 
oxygen covers both fermentative processes, the loss due to each 
process may be calculated with sufficient accuracy for practical 
purposes on the assumption that the loss due to the nitrogen 
fermentation is equal to the volume of oxygen theoretically required 
for the direct oxidation of the quantity of ammonia found to have 
been oxidized by fermentation. (I have fully treated this question 
in my original paper.) 
The samples 3, 4, and 5 in Table I. were all collected on Octo- 
ber 16th, and in dry weather. Their analyses again illustrate 
the very great value of dissolved gas determinations before and 
after fermentation. They indicate a very great difference in 
quality between samples 3 and 4 on the one hand, and sample 5 on 
the other. 
In sample 3, a slight fermentation took place on keeping, but 
it was practically confined to a nitrogen fermentation. In sample 
4 the fermentation was more decided, and it consisted of a slight 
carbon as well as a slight nitrogen fermentation. 
The analysis for sample 5 shows not only a very decided 
fermentation, but also that it was entirely confined to a carbon 
fermentation. As a matter of fact, 0°28 part of carbon per 
100,000 parts of the water underwent fermentation, and ‘033 part 
of nitrogen were converted into ammonia during the process, 
but it is to be noted that about one-third of this nitrogen was 
derived from nitrates originally present, and only two-thirds from 
unfermented organic matter. 
The analyses therefore show that the ratio of carbon to 
nitrogen in the unfermented organic matters originally present in 
the water was 1:10, a ratio which, taken together with the fact 
that the organic matters were unaccompanied by ammonia, 
proves them of vegetable origin, and that they were not derived 
from sewage. As a matter of fact I may explain that they were 
due to decaying vegetable matter of the autumn season. 
The analytical details for samples 6 and 7 further illustrate 
the valuable and definite results obtainable by the aération method. 
I need only refer to one point in connexion with them. Sample 
6 was collected on October 25th, when rain was commencing to 
1 Trans. R.D.S., vol. v., pp. 598-604. 
