90 Calculation of the horizontal Refraction 
the atom of nickel to weigh 30, and that of oxygen 8, hydrogen 
being unity) gives 10°20 per cent. for the contents of the speci- 
men analysed in metallic nickel. 
100 grains of titanious iron from North America, being dis- 
solved in muriatic acid (after the requisite ignition with potash), 
were treated (after separating the titanium) with excess of car- 
bonate of lime and filtered. ‘The excess of carbonic acid being 
expelled, ammonia was added, and a small quantity of a white 
precipitate fell, which speedily blackened in the air, and proved 
to be mere oxide of manganese, uncontaminated by iron, and 
amounting to half a grain. 
Manganese has been suspected in various species of cast iron; 
and though Mr. Mushet’s experiments go to prove that it does 
not usually enter in abundance, they can hardly be regarded as 
establishing the fact of its absence. It might not be uninterest- 
ing to resume the investigation with the aid of a mode of ana- 
lysis so well adapted to experiments on a large scale, as I have 
no doubt that, with proper care, one part ‘n a thousand, or even 
less, of manganese might be insulated from iron. 
The separation of iron from uranium cannot be accomplished 
by the process above described, that metal possessing a pro- 
perty analogous to that which forms the subject of this paper. 
By inverting the process, however, we shall succeed even here. 
A mixed solution of iron and uranium being deoxidized by a 
current of sulphuretted hydrogen, and then treated with an earthy 
carbonate, the iron passes in solution while the uranium separates. 
This difference in the habitudes of the two oxides of iron pre- 
sents us in fact with a kind of chemical dilemma, of one or the 
other of whose horns we may avail ourselves in any proposed 
ease. In studying the habitudes of uranium, however, I have 
met with some anomalies which require further investigation. 
Zirconia too might probably be freed from iron with equal fa- 
cility by a similar inversion of the process; but this I have not 
yet had an opportunity of trying satisfactorily. 
London, April 4, 1821. J. F. W. HerRscHeE.. 
XIX. Calculation of the horizontal Refraction in an Almo- 
sphere of uniform Temperature. By James Ivory, M.A. 
F.R.S, 
I; may not be ungratifying to some of the readers of the Philo- 
sophical Magazine, to have laid before them the method usually 
employed for computing the horizontal refraction in an atmo- 
sphere of uniform temperature, of which so much has lately been 
said. For this purpose, turn to the Magazine for September 
last, 
