282 



reasons platinum should have come last, under palladium. A re- 

 investigation of the atomic weights, suggested by these consider- 

 ations, was undertaken by Seubert, and the old values were 

 found in fact to be very inaccurate. He obtained: 



Os = 191, Ir = 193, Pt = 195. 



Originally lead, although it fell in the fourth column, possessed 

 only one compound Pb0 2 in which it seemed to be undoubtedly 

 quadrivalent. Search for salts of the same form, however, speed- 

 ily yielded the tetrachloride PbCU, tetracetate, and many others. 

 The existence of osmic acid Os04, and a corresponding compound 

 of ruthenium, suggests that other compounds of the elements of 

 the eighth group, displaying the valence eight, may be capable 

 of preparation. The collocation of copper, silver, and gold, in 

 the same column with the alkali metals, is not at present per- 

 fectly satisfactory, and suggests the advisability of strengthening 

 their position, if possible, by further investigation. 



In the same way, incorrect values of many physical properties 

 have been detected, and have been rectified by more careful work. 



4. By furnishing a comprehensive classification of the elements, 

 arranging them so as to exhibit the relationships among the physi- 

 cal and chemical properties of the elements themselves and of their 

 compounds. Constant use will be made of this property of the 

 table in the succeeding chapters. Having disposed of the halogen 

 and sulphur families, situated, respectively, in the eighth and 

 seventh columns of the table on p. 278, we shall presently take up 

 nitrogen and phosphorus from the right side of the sixth column. 

 Then from the fifth column, we shall select carbon and silicon, 

 and from the fourth boron, leaving the other, more decidedly 

 metallic elements for later treatment. 



Defects in the Periodic System. The periodic system is 

 often described as if it furnished a classification of the properties 



