PROCEEDINGS OF SECTION B. 327 



up the filter paper. It was then neutralised with ammonia in slight 

 excess, boiled, filtered, and washed free from any chlorides with am- 

 monium nitrate water. This Avas readily done, as but little chloride 

 remains present. The precipitate was then ignited, and weighed in 

 the crucible containing the residue from HF evaporation of the main 

 precipitate of silica. 



The results of the double duplicate for manganese are very satis- 

 factory, showing the ammonia separation to be quite satisfactory. 

 (See under " Manganese.") 



The weighed precipitate containing the ashes from two filters was 

 then fused vnth KH SO4, and filtered. The intermixture of the first 

 filter paper has the effect of keeping the precipitate finely divided, and 

 renders the fusion a very quick one (about 15-20 minutes at most). 

 In the ordinary course the FeoOg Al;j03 precipitate, Avhen ignited, forms 

 hard lumps, which require fusion for two or three hours. Formerly it 

 was our custom to mix the ignited oxides Anth KHSO^, and agate before 

 fusion ; but this is likely to be dangerous. The use of the filter paper 

 overcomes this difficulty of fusion, and no trouble has been experienced 

 in filtering. After ignition, the precipitate should be moistened with 

 strong nitric acid and reignited. Complete reoxidation may not be- 

 effected, but the amount of error is so small that when the iron is less 

 than 5 per cent, it may be neglected. With a high iron content it is 

 safer to ignite the precipitate separately from the paper. 



The melt from KHSO4 fusion is taken up with hot water and 5 

 per cent, to 10 per cent. H2SO4, which is necessary to prevent precipita- 

 tion of metatitanic acid. In this, as in all fusions, the greatest care 

 should be exercised lest the platinum be injured. It is convenient to 

 complete the fusion in the evening, and allow the crucible and melt to 

 soak over night. Removal of the melt, by alloAving it to solidify around 

 Pt. rod or wire is very simple. 



The silica recovered here is filtered, weighed, and evaporated with 

 HF, and the amount added to the weight of the main lot. If only 

 one evaporation has been used for SiOj, as suggested, this may amount 

 to -0050, or more, as shown in tables I. and II. 



Reduction of Iron. — Experiments have shown that reduction 

 of the ferric iron by amalgamated zinc in contact with platinum gives 

 almost as good results as reduction by HjS. The zinc must, of course, 

 be free from iron, and can generally be obtained so now. The objections 

 to the zinc reduction are that vanadium, if present, is reduced to V2O3, 

 and so will count as iron in the permanganate titration, and that 

 titanium sulphate is also reduced, and the subsequent determination 

 of that element interfered with. 



If vanadium is present it will be reduced in either case, though 

 the HjS method will only produce the condition of V2O4, whilst Avith 

 nascent hydrogen it AA'ill partially, if not entirely, reduce to V2O3 (Hille- 

 brand). 



Experiments were made with each method as foUoAA's : — Standard 

 solutions of ferric iron (A) and titanium sulphate (B) Avere prepared, 

 and titrated with standard KMnO^ (Ice. = -0050 Fe.). 



