102 
ON THE DOUBLE SALTS OP IRON. 
of inductive science ; I determined to rest the preceding part 
of my argument on what, with all the correction I have be- 
stowed on it, still remains weak and defective, because, 
trusting to my formula, I had given to several gentlemen 
samples of a salt which proved to be less perfect than I 
have made it at other times. This difficulty, which op- 
poses itself to the establishment of a superseries among ci- 
tric salts, will not be felt when I come to treat of the ana- 
logous compounds with ' tartaric acid. We have been long 
familiar with an example of the inferior series of tartaric 
acid salts in the officinal potassio-tartrate of iron ; I there- 
fore shall not now further occupy myself with them. The 
superseries is made evident in the following manner: — The 
bitartrate of iron resulting from the operation in case 4 is to 
be saturated with potash or ammonia ; the solution is to be 
then filtered and carefully*dried in small quantities at a time; 
I found ammoniacal salt to contain thirty-five grains of oxide: 
estimated theoretically, it is a fraction short of this quantity. 
Thus- 
One equivalent of oxide of iron, 80 or 34.93 
Two equivalents of tartaric acid, 132 " 57.64 
One equivalent of ammonia, 17 " 7.43 
229 100.00 
To one or other of these two series, or to a mixture of 
both, the whole of the double iron salts, containing the ses- 
quioxide, with which I am acquainted, are capable of being 
referred. The difficulty attending the manufacture of per- 
fect salts of the superseries, is such as to render it improba- 
ble that they can ever become extensively used. The am- 
monio-citrate of commerce contains an average of thirty 
per cent, of oxide ; and in this state it is more permanent. 
A salt having this constitution is made without much trouble 
as follows:* 
* The preparation of the hydrated oxide is an important point in the 
formation of these salts ; the London College orders the sesquichloride 
to be decomposed with caustic potash ; the Edinburgh directs the per- 
