438 Pwceedincjs of Royal Society of Edinburgh, [july 16 , 
By chemical and microscopical investigations afterwards I found 
that this disappearance of the acid reaction was undoubtedly due to 
the absorption and combination of the carbonic acid gas held in 
solution by and with the magnesia especially, but also, although 
in a much less degree, with the protoxide of iron. 
After the olivine had lain undisturbed in the carbonated water 
for two months, I took a small portion of the liquid and tested it 
directly for magnesia as follows : — I added to it first about a fourth 
of its bulk of ammonium chloride, and after thoroughly mixing saw 
that the liquid remained perfectly clear; I then poured in a sufficient 
quantity of ammonium hydrate to make the whole fluid, after again 
mixing properly, distinctly alkaline. To the still perfectly clear 
liquid I now added ammonium phosphate, and shook up the mixture 
very briskly for two or three minutes. A very distinct white 
crystalline precipitate of the double phosj^hate of ammonium and 
magnesium formed almost at once, proving the presenceof magnesium, 
and showing that that metal had been removed from the olivine 
crystals by the action of the carbonic acid water. 
I tested another small portion of the water in which the crystals 
had lain, for iron, by the following methods : — (1) I poured in a 
little rather strong pure hydrochloric acid (free from even a trace of 
iron) heated gently, and then added in small quantities at a time a 
little solid chlorate of potash,* and continued heating for four or 
five minutes. By this operation any iron present in the ferrous 
state was changed into the ferric condition, and rendered fit for 
testing. To one part of this ferricdsed solution I now added 
thiocyanate of potassium, when a pale but distinct blood-red 
coloration was produced, showing that a trace of iron had been 
removed from the olivine, and held in solution by the carbonated 
water. (2) I also converted, in another portion of the original 
solution, the ferrous salt present into ferric, by boiling it for some 
time pure concentrated nitric acid (absolutely free from iron), 
and tested the liquid so treated by the methods detailed above for 
magnesia and iron, with precisely the same results. 
After removing the crystals I evaporated off the remainder of the 
liquid, in which they had been placed, over the water-bath, and 
I used the chlorate of potassium, but am now aware that potassium 
permanganate would have served the purpose better. 
