326 



PROCEEDINGS OF SECTION B. 



As in the previous table, the analyses in which the silica received 

 a single evaporation for SiOj are marked A ; those marked B receiving 



two. 



Table IT. 



Silica SiO^. 



Altjmtna, Iron, Etc. 



A— 1 



B— 1 



Combined 



Al203,Feo03 



P„6j ppt. 



.1620 

 .1627 



.156« 

 .1593 



Correction 



(Volatile 

 SiO,). 



.0078 

 .0075 



.0027 

 .0027 



.1548 

 .1542 



,1541 

 , 1566 



Basic acetate reparation 



These tests were made with every care for each method, and the 

 comparison is interesting. In both silica and alumina the duplicate A 

 figures (single evaporation) are nearer one another than in the B 

 duplicates (double evaporation). The final SiOg for A is higher, and 

 in the alumina only 2B is higher, but this one shows a serious discrepancy 

 with its duplicate. In B, the basic acetate separation had been used, 

 but filtration was slow. The extra manipulation involved probably 

 accounts for the poorer figures. 



Precipitation of Alumina, Iron, &c. — Hillebrand advises at 

 least one precipitation by basic acetate if any manganese is present at 

 all, but admits that two ammonia precipitations are usually quite 

 satisfactory for complete separation of AI2O3, FegOa, P20g, &c., from 

 the manganese. There is no doubt that the use of basic acetate is apt 

 to give incomplete precipitation of alumina, and it is always liable to 

 serious trouble in filtering, even when the greatest care is exercised. 



In experiments tried on this point the filtrate from the silica was 

 boiled, ammoniam chloride (25ccs. saturated sol.) added, then ammonia 

 in slight excess, the solution heated for five minutes and filtered, washed 

 twice with ammonium nitrate sol. (2 per cent.), dissolved in HCl, 

 reprecipitated and filtered. The precipitate and paper was then re- 

 turned to the beaker, in which precipitation took place, dissolved in 

 HNO3 (5 CCS. fort) diluted to 50 ccs., warmed and stirred well to break 



