51 
obtained were identical with those already cited. In two 
of these determinations caustic potash was substituted for 
ammonia: the results obtained deserve mention: — 
TT wt. of 
x± ‘ Fe(NH 4 ) 2 
2S0 4 '6H 2 0. 
Per centage 
Fe 
theoretical. 
Wt. of 
KOH&NaCl 
added. 
Per centage 
Fe fromFeoOa 
obtained. 
, Error. Note. 
per cent. 
9- 
•706 
14-28 
Grms, 
[5 KOH 
1 5 NaCl 
17-93 
Wash watei~950 cc. 
pptd. iu boiling so- 
° ^ ° lution 
k 
•720 
• • • 
V 
• • • 
18-76 
4 '48 Wash water=400 cc. 
pptd. in boiling so- 
The precipitates were redissolved in concentrated HC1 
after weighing. They were observed to effervesce violently 
on treatment with the acid, and on evaporating the solution 
the alkaline chloride crystallised out in considerable quan- 
tities. 
I instance these as indicating the extent of adhesion 
existing between ferric oxide and the fixed alkalis, and the 
futility of the practice, which I believe is not unusual, of 
weighing a precipitate of the oxide from which alumina ha s 
been removed by means of potash, without previously re- 
dissolving and again precipitating. 
Determinations were then made in which special precau- 
tions were taken with the view of ascertaining the limit of 
the error due to the retention of alkaline salt. A weighed 
quantity of pure iron was taken for each experiment. The 
results are appended : — 
III. 
Wt Fe calc. 
Wt. Fe 
NaCl 
from Fe ff 0 3 
Error. 
Note. 
taken. 
added. 
obtained. 
k 
T669 
o-ooo 
•1666 
—0-00031 
\ Washed persistently 
l 
•2088 
1-000 
•2096 
+ 0-0008 
[ with boiling 
till washing 
water 
abso- 
m. 
•1485 
1000 
•1498 
+ 00013 j 
l lately free 
1 Chlorine. 
from 
The impurity present in the precipitate in m was isolated 
in the manner described above and weighed. It was found 
= *0023 grm. Multiplying this by 0'7, the fraction ex- 
pressing the ratio Fe 2 :Fe 2 0 3 , we obtain the number 00016, 
which it will be seen is in close agreement with the error 
obtained (calculated as Fe). 
In the following experiment pure iron was brought into 
solution as Fe 2 Cl 6 , and precipitated by ammonia alone. 
