[M‘GILz] OXIDATION OF FERROUS CHLORIDE 49 
First Series.—On examining the work of this series in the light 
of later results, it is evident that the disappointing record shown is 
chiefly due to the small percentage of water still present in the sam- 
ples of ferrous chloride used. I had taken, as I thought, all necessary 
precautions to have both the chloride and oxide thoroughly dry. 
The influence of mere traces of water in the charge, is very 
perceptible in destroying available chlorine, and is expressed by the 
formula = x 10 — _ per cent of chlorine in the sample, 
à a 
where :— 
a — per cent chlorine in sample. 
es NT water “ 
Hence, the loss of free chlorine for each unit per cent of water in 
Sample I. is 10-8 per cent, and in Sample III. is 20-2 per cent of 
the total chlorine in these samples, respectively. 
The highest result in free chlorine obtained with Sample I. is 
33-1 per cent (Expt. 16), and this is also the best yield of the series. 
Taken in connection with other experiments in which Sample I. 
was used, it shows the favourable influence of a large proportion of 
ferric oxide, a condition fully confirmed by later results. 
Especial pains were taken to dry Sample III., yet the best result 
obtained with it was 20 per cent available (Expt. 9), and this although 
in Expt. 12,as much as 91-6 per cent of its total chlorine passed over 
into the absorbers. The fact that for each 1 per cent of water 
present, 20-2 per cent of chlorine reverted, was not taken into con- 
sideration at the time, and, in consequence, the investigation was 
discontinued during an interval of nearly nine years. 
Second Series.—The investigation was resumed in December, 1899, 
chiefly on account of the fact that, during the interval, methods for 
the production of oxygen (or highly enriched air) had been developed, 
and it was believed that oxygen could be economically employed for 
the oxidation of ferrous chloride. The samples used were dried with no 
greater care than those employed in the First Series, but Sample V. must 
have been very dry, since in several experiments it yielded about 60 
per cent of its total chlorine in the free state. It is noteworthy that 
in one of these experiments (No. 15) air only was used; and the chief 
positive result of this series is to demonstrate that, so far as liberation 
of chlorine is concerned, oxygen has no important advantage over 
atmospheric air. Perhaps other considerations, such as the dilution 
of the product by atmospheric nitrogen, may be found to give cheap 
oxygen an advantage over air; but the examination of this point does 
