1892.] On certain Ternary Alloys. 13 



considerable, small amounts of iron and silicon were also contained 

 therein. 



In some cases, when the amount of aluminium in the upper alloy 

 was large and the other constituents small, the aluminium was taken 

 by difference, the percentages being corrected by means of the above 

 analysis to what they would have been if reckoned on the sum of the 

 aluminium and the other two metals as 100. When, however, the 

 other constituents were present in larger quantity the aluminium was 

 directly determined, the alumina ultimately precipitated being col- 

 lected and weighed, and the Si0 2 and Fe 2 3 present therein subse- 

 quently determined and subtracted so as to obtain a corrected deter- 

 mination of the aluminium ; the percentages were then reckoned on 

 the sum of the aluminium thus found and the other two metals as 

 100. Thus, for example, an alloy of tin, lead, and aluminium was 

 found to contain 12'01 per cent, of tin and 1*71 per cent, of lead; 

 hence, reckoning the difference, 86'28 per cent., as aluminium, silicon, 

 and iron containing 95'81 per cent, of the first, the corrected ana- 

 lysis is 



Tin 12-01 = 12-46 



Lead 1-71 = 1-77 



Aluminium. . ,86'28 x 0*9581 = 82'66 = 85'77 



96-38 100-00 



On the other hand, an alloy of silver, aluminium, and lead was 

 found to contain the following percentages of these three metals 

 respectively, the aluminium being reckoned from the weight of 

 alumina after subtraction of silica and ferric oxide contained in the 

 mixed precipitate first weighed ; whence the annexed composition, 

 calculated on the sum of the three metals as 100 : 



Calculated on sum 

 Found. as 100. 



Silver 80'17 8079 



Lead 6'55 6'60 



Aluminium 12*51 12'61 



99-23 100-00 



In the case of the lighter alloys containing tin, aluminium, and 

 lead (or bismuth) with large percentages of the first metal, it was 

 sometimes found that a notable amount of oxygen was absorbed by 

 the ingot whilst standing molten for eight hours, so that a per- 

 ceptible deficiency from 100 was observed when all the constituents 

 were added together. With silver-aluminium-lead and silver- 

 aluminium-bismuth alloys this was not the case ; on the other hand, 



