330 



PROCEEDINGS OP SECTION B. 



Phosphorus. — Hillebrand requires a fusion with alkali carbonate 

 and removal of silica exactly as in the principal determination of SiOg, 

 except that nitric acid is used instead of HCl. Washington proposes, 

 instead, to break up with nitric and hydrofluoric acids in platinum, to 

 remove SiOa by evaporation and filtering, and to precipitate the P2O5 

 as molybdate in the usual manner. 



Although no comparative tests have been made, good results have 

 been obtained by careful digestion with aqua regia, evaporation to dry- 

 ness with HNO3, and direct precipitation in the filtrate with molybdate. 



The more severe treatment is desirable, according to Washington, 

 on account of inclusions of apatite, which would not be attacked by 

 the ordinary treatment. If the preliminary agating of the sample is 

 carefully done, this objection does not appear to hold. This point is 

 under further investigation. Hillebrand does not agate his sample in 

 preparation, and insists on a fusion. 



Ferrous Iron. — Instead of the water-bath arrangement of Cooke, 

 excellent results have been obtained by heating one gramme of sample 

 in a large platinum crucible, with a well- fitting cover, with sulphmic 

 acid (1-1) and hydrofluoric acid on a hot plate. After 10 minutes the 

 crucible and contents are quietly dropped into a beaker of cold water, 

 and titrated with Kmn04 (1 cc. = -0050 Fe). With care, perfectly 

 satisfactory results are obtained without using Coj. Thus, using one 

 gramme of material with 10 ccs. H2SO4 (1-1) and 5 ccs. HF. 



Tablt? VI. 



Moisture : Loss by Ignition. — Experiments were made to test 

 the error involved by the " loss on ignition " method, compared with 

 Penfield's method. Hillebrand points out that it is unsafe to apply a 

 correction on the assumption that all ferrous iron present has been 

 oxidised. In the experiments made, this was not taken for granted ; 

 but, directly the crucible was weighed after ignition, the actual ferrous 

 iron remaining was estimated as above shown, and the correction ap- 

 plied. The material worked on was a sandstone containing about 10 

 per cent. CaCOg. The complete analysis is shown below. One gramme 

 taken in a weighed platinum crucible and dried at 110° C. in air-bath 

 till constant. Loss = moisture below 110° C(H20-). 



The crucible was then treated over a Bunsen flame till weight was 

 constant = total loss on ignition. The ferrous iron remaining was 

 then quicklv determined by moistening with water and treating with 

 H2SO4 and HF, and titrating with permanganate. An amount equiva- 



