344 Dr Thomson's Observations on the Weights of [Nov. 



Weight of an atom. 



27 Glucinum 2-250 



28 Glucina 3-250 p 



29 Aluminum 1-125 



30 Alumina 2-125 i 



31 Zirconium 4-625 ? 



32 Zirconia 5-625 1 



33 Iron 3-500 



34 Protoxide of iron 4-500 r 



35 Peroxide of iron 10-000 s 



36 Nickel 3-375 



■ 37 Protoxide of nickel 4-375 * 



+ 1 atom oxygen ; but we have no means of determining whether 

 the supposition be well or ill founded. It would be necessary to 

 determine the weight of oxygen and metal in yttria before we 

 could be quite certain. 



p According to Berzelius, sulphate of glucina is a compound of 



Acid. Glucina. 



100 + 64-100 

 If it be composed of 5 acid + 3-25 glucina, its 7 -,qq , gc.AQQ 

 composition will be 5 



i This number is founded upon Berzelius's analysis of sulphate 

 of alumina. — (Ann. de Chim. lxxxii. 14.) 



Acid. Alumina. 



He found it a compound of. 100 + 42-722 



Now 100 : 42-722 :: 5 : 2-115. I take 2-125 as sensibly the 

 same with 2-115. Mr. Richard Philips informs me that he has 

 analysed the sulphate of alumina with a different result; that the 

 quantity of alumina which he found in the salt was much greater. 

 Should this statement turn out correct, the weight of an atom of 

 alumina would be higher than it is stated above, and might even 

 amount to 3-5, the number in my original table, 



r According to Berzelius's analysis of sulphate of iron, the 

 weight of protoxide of iron is 4-4. He found it a compound of 

 28-9 acid + 25-7 protoxide of iron + 45-4 water. Now 28*9 : 

 25*7 :: 5 : 4*4. I have adopted 4-5 as sufficiently near the 

 number. The mean of all the analyses of protoxide of iron 

 gives us 100 iron + 28-78 oxygen. If we consider protoxide of 

 iron as a compound of 1 atom iron + 1 atom oxygen, and its 

 weight to be 4-5, then it will be a compound of 100 iron -+• 

 28*57 oxygen. These coincidences are sufficiently near. 



s My reasons for pitching upon this number for the weight of 

 an atom of peroxide of iron have been given in the Annals of 

 Philosophy , x. 98. 



J I am not quite satisfied with this number, though it is 



