ELISHA MITCHELL SCIENTIFIC SOCIETY. 15 
The larger portion of the iron may be separated by the 
use of sulphur dioxide. The impure zirconium chloride 
is made up into a moderately dilute soiution in a large 
jar. This is nearly neutralized with ammonia and sul- 
phur dioxide is passed through. The zirconium is pre- 
cipitated out as a basic sulphite. This is filtered away 
from the solution containing most of the iron. In most 
cases it is advisable to dissolve again in hydrochloric acid 
and precipitate once more with sulphur dioxide. The 
filtrate containing the iron also contains much of the zir- 
conium. A good deal of this is precipitated on boiling. 
The filtrate from this is then evaporated toa small bulk 
and a good deal of the iron will crystallize out as the dou- 
ble chloride of iron and ammonium in deep red crystals. 
The concentrated filtrates from several fusions, thus part- 
ly freed from iron, may be again treated with sulphur 
dioxide and the zirconium recovered. 
The sulphite after draining in large funnels is dissolved 
in hydrochloric acid and boiled until all sulphur dioxide 
is driven off. It is then precipitated as hydroxide by 
means of ammonium hydroxide and well washed. This 
hydroxide is drained upon filters and dissolved in the 
least amount of concentrated hydrochloric acid. The 
remaining impurities are gotton rid of by repeated 
crystallizations from strong hydrochloric acid. This 
must be managed with some care as it is quite easy to 
form an oxychloride insoluble in the strong acid, for in- 
stance by the addition of strong acid to a somewhat concen- 
trated aqueous solution. [In evaporating the hydrochloric 
acid solution to crystallization it is therefore necessary to 
add concentrated acid from time to time as the watery 
acid evaporates so that the first crystallization shall be 
from strong acid. After that there is little trouble in 
subsequent crystallizations from strong acid. The crys- 
tallizations are best made from a casserole and one hun- 
dred grams or more of chloride may be crystallized at 
atime. Generally twelve or fifteen crystallizations will 
