186 
Proceedings of the Royal Irish Academy. 
the typical salts, and adding degree by degree to them, whilst in a 
state of ebullition, water, until the thermanalytic point was reached ; 
they were so arranged that the evaporating water was recondensed, 
and flowed back into the flasks. The water was distilled twice, the 
first time from a small quantity of acid- sulphate of potassium, to render 
it free from all trace of ammonia. The flasks were made of hard Ger- 
man glass, and were digested with dilute hydrochloric acid for some 
time before they were used. It was found impossible to obtain water 
by filtration through paper that would stand the test of a beam of 
the electric or lime light. The clearest water was obtained by subsi- 
dence after distillation from a silver retort. Except in a few cases, 
most of the experiments can be followed by performing them by gas- 
light. The flasks must, in this case, be carefully cleaned, and a power- 
ful gas-flame be placed opposite the manipulator and behind the flask. 
The faintest precipitate will be perceived by this means, on holding 
the flask close to the eye, providing it is not of that peculiar transpa- 
rent and colloidal nature sometimes observed in modifications of alu- 
mina and silica. 
The solutions used were weak, but were so proportioned that each 
contained an equivalent quantity of the trioxide to correspond to one 
gramme of chloride of aluminium, dissolved in the half litre of water. 
The exceptions were in two of the iron experiments. The following 
were the results : — 
Temperature 
at which, dis- 
sociation sets 
in. 
Degrees, cubic 
cent., of dilution 
required to pro- 
duce decompo- 
sition with 1 
gramme Al 4 Clj 
or its equiva- 
lent. 
53-5° C. 
9-5 C. C. 
J 48 
50° C. 
8- C. C. 
43 
42 
55001 C. C. 
2751 C. C. 
66251 C. C. 
31260 C. C. 
Experi- 
ment. 
22 
23 
24 
25 
26 
27 
28 
29 
30 
31 
Salt used. 
FeNH4 2S04,12H20. 
FeK2S0i, I2H3O. 
CrK2S04,12H20. 
AINH4, 2S04,12H20. 
Weight used, or equi- 
valent to one gramme 
of AI2CI4 in 1 litre H^O. 
3-597 
( With one vol. 
( more H2O. 
( With two vol 
\ more H2O. 
3-753 
With one vol 
„ two „ 
1-186 
3-727 
1 • 
3-384 
H2O. 
The observations recorded in these experiments can only be consi- 
dered exact from a relative point of view, as there were too many 
contending influences to warrant the figures being taken as individual 
observations. They, however, conclusively prove the following points : 
The relative position that the molecules, ferric, chromic, and aluminic 
