REPORT OF CHEMIST. ~ 1095 
loss in weight. The constancy, in this respect, was tested by aspirating 
a considerable quantity of air through them, and weighing repeatedly. 
Sulphur was determined by dissolving in hydrochloric acid, and passing’ 
the resultant gas through U tubes, containing a solution of potassic 
permanganate (Drown’s method). The solution in the flask was then 
filtered, the residues fused with a little nitrate and carbonate of soda, 
dissolved, evaporated with acid and added to the permanganate solution 
After reducing, the clear solution was precipitated by baric chloride. 
Five grammes of the iron was used for each determination. The iron 
was in most cases prepared for analysis by drilling with a dry, clean 
drill, and mixing the chips from several parts of the sample. Silicon 
was determined by Drown’s nitric and sulphuric acid method. So much 
discussion has taken place as to the determination of phosphorus, and, 
as quite a number of results are given in the chapter on iron which are 
taken from private notes of former work and will not be repeated in 
the tables of analyses added here which only include those made -for 
the survey and which results were obtained ‘by the same method, it 
will be given in full. 
Two grammes of the iron were placed in a beaker, with 200 c. ¢. 
of water and 4 or 5 grammes of chlorate of potash, the whole heated 
until the salt was dissolved, and hydrochloric acid was then added 
gradually until the iron was dissolved. By taking time, the solution 
takes place without any notable evolution of gas. 
The solution ‘so obtained was evaporated, dried, and taken up with 
hydrochloric acid. The residue from filtration was fused with sodic 
nitrate and carbonate, and the fused mass tfeated as in iron ores, the 
solutions being combined. The liquid then was treated as in the case of 
iron ores. The fusion of the residue seems to be absolutely necessary 
to obtain all the phosphorus, as from 75 to zy of the whole present has 
been shown by numerous trials to be retained there after the first solu- 
tion where silica is present to any considerable extent. The same 
results are obtained if nitric acid is used as the original solvent. 
The above method is essentially the one used for years at the School 
of Mines in New York, and published by Mr. Cairns in his book on 
analysis ; it is, according to my experience, the most satisfactory, though 
not the shortest. It seems that no general method is satisfactory for 
phosphorus, that omits the examination of every product for phosphorie 
acid; such a method as will give uniform results with materials similar 
