3824 REPORTS ON THE STATE OF SCIENCE. 
specially hardened material and the sample of rock so big, the chance of 
contamination is very slight, whereas the saving of time is enormous. 
The powder thus obtained is sufficiently fine for the main fusion, but 
for the determination of ferrous oxide and alkalies the rock must be in 
a much finer state of division. For this purpose some of the main stock 
of powder was taken and ground by hand in an agate mortar until it was 
sufficiently fine to pass through a 120-mesh sieve. 
Apparatus and Chemicals.—All fusions were carried out in platinum 
crucibles, platium dishes being also employed for all evaporations ; a large 
silver basin was used in the estimation of alkalies. The glass beakers and 
flasks were all of a special make (R glass), very resistant to the action of 
solvents. The pipettes, burette and other measuring instruments were all 
standardised. 
The reagents used were Merck’s (specially pure); these were first 
carefully tested for impurities, and corrections applied for them where 
necessary. One sample of hydrofluoric acid contained an appreciable 
amount of iron, which had to be corrected for, and the calcium carbonate 
contained alkalies. The amount of alkalies present was determined in 
20 grams of the carbonate, and the values obtained were used to appl 
corrections in the determination of alkalies in the rocks. Freshly distilled 
water was used throughout. 
Methods.— Estimation of SiO,, Al,O3, Fe,O;, MnO, CaO, MgO. For 
the determination of these about a gram of rock-powder was fused with 
about six times its weight of fusion mixture until decomposition was 
complete. From this fused mass the silica was separated and esti- 
mated. The metals were converted into chlorides and precipitated 
from solution in their proper order, iron and aluminium as hydroxides, the 
Fe,0, estimated by standard permanganate after reduction with H,S, and 
the Al,O, determined by difference ; manganese was precipitated as 
sulphide, calcium as oxalate, with subsequent conversion into oxide, 
magnesium as phosphate, the precipitates in all cases being dissolved and 
reprecipitated to ensure their purity. 
Ferrous oxide was determined in about half a gram of the specially 
ground powder by solution in a mixture of sulphuric and hydrofluoric 
acids at a boiling temperature and immediate titration with standard 
permanganate solution. 
Alkalies—For the determination of alkalies the Lawrence Smith 
method was adopted. 
The method was first tested by making a determination with a weighed 
quantity of pure potassium and sodium chlorides, and it was found to be 
quite reliable. 
About half a gram of the specially ground powder was intimately 
mixed with ammonium chloride and calcium carbonate, and fused. The 
alkalies are thus converted into chlorides. The mass was thoroughly 
leached with water, and, after separating all traces of ammonium chloride 
and calcium carbonate, the mixed alkali chlorides were weighed. The 
potassium was separated as platinichloride and weighed, and the sodium 
determined by difference. A correction had to be applied for the alkalies 
present in the weighed quantity of calcium carbonate taken. 
For the determination of CO, a weighed quantity of rock-powder was 
treated with hydrochloric acid, and the CO,, after thorough drying and 
purifying, absorbed in a Yj tube containing soda lime, precautions being 
taken to keep the whole apparatus full of a current of air free from CO,. 
