50 ROYAL SOCIETY OF CANADA 
not fall within the scope of the present investigation. (See Expt. 32, 
of Series III.) 
Third Series.—This series was undertaken in order to ascertain 
whether it was practicable to make ferrous chloride sufficiently dry, 
by roasting on iron plates, to obtain from it a high yield of free 
chlorine. Roasting in air is attended by more or less oxidation and loss 
of chlorine; so that experimental proof was wanted of the possibility of 
making the charge absolutely dry, without too great waste of chlorine 
at the same time. Samples A and B, used in this series, are identical 
with V. and VI. of the Second Series. The remaining samples are 
mostly prepared from these two, by mixing with ferric oxide or ferric 
hydrate, and drying. It is unfortunate that a galvanized iron plate 
was used in drying samples A and B, hence introducing chloride of 
zinc into the charge. The influence of this impurity is slight in sam- 
ples C, D, F and G, 
The importance of thorough drying had been seen, and the yield 
of available chlorine is much better throughout the whole series. The 
catalytic influence of Fe,0, is recognized, and the presence of a large 
excess of ferric oxide was found to be favourable to the preliminary 
drying of ferrous chloride. Samples F and G with which the work 
of this series was begun, contain about 65 per cent of ferric oxide; 
and sample CC, which gave encouraging results in experiments 12 to 
26, contained about 10 per cent of ferric oxide, and gave decidedly 
higher results in free chlorine when ferric oxide was added to the 
charge — (Expt. 18, 22, 25, 26). The fact that Fe,O, acts cataly- 
tically, and not merely mechanically is further borne out by the 
results of Experiments 21 and 24 in which dry gypsum and dry sand, 
respectively, were substituted for ferric oxide, with an immediate 
decrease in the percentage of available chlorine. 
Experiment 23 is introduced for incidental investigation of the 
question of recovery of chlorine from Solvay process waste, by a 
variation of this mode of working. The question is important 
enough to justify separate treatment; and there can be no doubt 
that special apparatus must be devised to overcome the difficulties 
introduced by the extreme hygroscopicity of chloride of calcium. 
The experiments made with sample EE (Nos. 28 to 33) prove the 
practicability of recovering the chlorine from Solvay waste by the 
use of ferrous sulphate. This sample contained about 15 per cent 
of calcium sulphate, yet gave a high yield of available chlorine. 
The final results of this work may be summed up as follows: — 
First.—That ferrous chloride can be decomposed by oxygen in 
such a way as to yield uniformly, from 75 to 85 per cent of its 
