444 . PROFESSOR HEDDLE ON 
then added and rubbed up. The grinding end of the pestle was then washed 
clean by a jet of water into the mortar, which in the so doing was now nearly 
filled. It was allowed to stand undisturbed for three minutes, when about two 
ounces of the muddy liquid was drawn off steadily by a pipette, and allowed 
to fall into a large precipitating glass, containing about thirty ounces of dis- 
tilled water. This was left undisturbed for ten minutes, when its contents, all 
but about two ounces at the bottom, were poured into a second larger preci- 
pitating glass. This was again left undisturbed for a quarter of an hour, when 
all, but about three ounces, was poured into a capacious glass jar. 
The coarser portions at the bottom of all the glasses employed were in turn 
returned to the mortar, and the process was continued and repeated, wntil every 
- portion of the quantity originally placed in the mortar was floated off, and uni- 
formly mixed in the one large settling jar. This was found to be absolutely 
necessary in some cases, (¢.g., in the magnetic sands from Granton), as some 
portions— either where there might be an admixture of ordinary magnetite, or 
some softening through incipient alteration into martite—were found to be 
much more readily comminuted than others. Such softer portions were found 
to contain less titanium, and more ferric oxide. 
The settling was generally complete in three days. 
Notwithstanding this extreme amount of subdivision, several of the sub- 
stances examined partially resisted decomposition by the ordinary methods of 
fusion with Fresenius’s flux,—potassium bisulphate,—and calcium and ammonium 
fluorides,—used singly, or even successively. 
It was, where possible, found better to operate upon an entirely new 
quantity,—comminuting and floating off still more finely,—than to recomminute 
the unresolved portion (mixed up with some flux to prevent loss). It was 
observed that the quantity which had jist escaped decomposition was more 
difficult to resolve even when recomminuted, than it was when fused up along 
with a quantity which was undergoing decomposition. 
This is an illustration of “ communication of energy,” similar to silver im- 
parting, in an alloy with platinum, the power of combination with nitric acid to 
the more noble metal. 
In several cases of fusion with potassium bisulphate, the separation of the 
titanic acid was found to be either slow or incomplete, some of it coming down 
at later periods of the analysis. In such cases the following process, somewhat 
modified from one recommended in a foreign journal, was adopted. 
After fusion with Fresenius’s flux and solution in acidulated water, with 
separation of the silicic acid and some titanic acid, ammonium chloride in 
strong solution was added, and then ammonia in slight excess. 
The precipitate of ferric oxide, alumina, and titanic acid, thus thrown down, 
was filtered off, washed, ignited, and weighed. It was then mixed with potas- 
