1813.] Definite Proportion* in Chemical Comlinations. 47 



Number of Weight of an 



atoms. integrant uarticlr, 



34. Potash 1 p + 1 o 6-000 



35. Peroxide of potash 1 p + 3 o 8-000 



36. Soda 1 s + 2 o 7*882 



37. Peroxide of soda Is + 3o S'882 



38. Barytes 1 b + 1 1)731 



39. Strontian 1 st + 1 6-900 



40. Lime .. .-I I + lo 3'620 



41. Meenesia 1 m + 1 o 2-368 



42. Alumina 2-136 s 



43. Glucina 3-600 



44. Yttvia ; 8-400 



45. Zirconia 5"656 



46. Silica 4-066 ' 



47. Gold 24-968 » 



48. Platinum 12-161 x 



of the alkaline earths ; but they stand upon rather better evi- 

 dence, though it is not quite satisfactory. 



' This, arid the three numbers below it, though founded on 

 the best analyses of the salts of these respective earths hitherto 

 made, ought not to be. much depended on, especially the number 

 for yttria, which is derived from Klaproth's analysis of the car- 

 bonate of yttria. Of all the genera of salts the carbonates are 

 the least to be trusted in such cases ; because the affinity of the 

 carbonic acid for the base is so weak that the salt cannot be dried 

 without losing a considerable portion of its acid. 



1 This number is derived from Mr. John Davy's analysis of 

 lilicated fluoric acid, and is not much to be depended on. I 

 have inserted it, however, that the numbers for the earths may 

 be complete. They will be found of considerable importance in 

 examining the different analyses of stony bodies hitherto pub- 

 lished. 



u This is derived from the experiments of Berzelius on the 

 oxides of gold, fie obtained 2 oxides; the first composed of 

 100 gold + 4-005 oxygen ; the second, of 100 gold + 11-982 

 oxygen. Now the second of these quantities is triple the other, 

 lie nee the first is a protoxide, and the second a peroxide. 



x This number is also derived from the experiments of Ber- 

 zelius. He obtained 2 oxides of platinum, the one composed 

 100 metal + 8-2S7 oxygen, the second of 100 metal + 

 16*36 oxygen. Heme we arc warranted to consider the first as 

 B protoxide. 1 have taken the mean of the two results, which 

 8*223 for the oxygen of the protoxide. He obtained the 

 first oxide by precipitating fresh muriate of platinum with 



