PROCEEDINGS OP SECTION B. 

 Table I. — continued. 



325 



Of the eight pairs of analyses, the amount of silica ultimately re- 

 covered was in six cases higher when two evaporations were used (Nos. 

 3 to 8), and in two cases lower (Nos. 1 and 2). 



Of the six higher results, three cases showed an advantage of under 

 one milligramme — two being almost identical, the excess being only 

 •0002 and -0005 grammes. One (No. 8) was higher by -0063 gramme, 

 but this result may safely be rejected, as the SiO.2 recovered from the 

 AI2O3 ppt. in this case (No. 8a) was unusually low (only -0006 gramme), 

 and the summation of the whole analysis was bad (99-10 per cent.). 

 The result is probably in error. 



If the corresponding alumina results are compared, it is seen that 

 in each case the use of double evaporation gave slightly higher results. 

 It would certainly appear that this treatment does tend to render the 

 AI2O3 more insoluble when it is finally precipitated. 



The above figm-es were not picked out, but taken in consecutive 

 order from a series of analyses of rocks from Macedon, Victoria, made 

 at the Mines Department laboratory by two chemists working in- 

 dependently and doing only single analyses. The personal factor, I 

 think, largely accounts for the higher AI2O3 figures, as,- whichever 

 method he adopted, one man nearly always obtained a higher figure 

 for alumina. 



The following example given is an analysis made on a rock by one 

 chemist (Mr. Alan G. Hall), when the personal element should be con- 

 stant. Each method, single and double evaporation, was made in 

 duplicate, and the results for silica and alumina show no advantage 

 from the second evaporation. The advantage of omitting this is not 

 only a saving of time, but also a removal of the danger of contamination 

 which will always accompany a long evaporation. This is especially 

 the case where economy does not permit the unlimited use of large 

 platinum vessels. 



