BETAVEEN THE ATOMIC 'WEIGHTS. 249 



This series may be naturally divided into two sub-series. The first contains tliose 

 elements whose protoxide bases are their characteristic compounds, and which do not 

 form acids with Oxygen. The second contains those elements whose characteristic 

 compounds are their sesquibases. They generally unite with two or more equivalents 

 of Oxygen, and form acids. These sub-series are distinguished in the table in the same 

 way as those of the Six Series. Corresponding to these sub-series we have two sets of 

 atomic weights, each having the same common difference, but differing in their starting- 

 point or nucleus. The first set is expressed by the formula 4 + « 4, the second by 

 2 + n4. 



The sub-series affiliate with each other in a most remarkable manner. Manganese, 

 for example, not only forms a strong protoxide base, but also unites with a larger 

 amount of Oxygen, forming both a sesquibase and acids. Its atomic weight places it 

 in the first group, and it has therefore been classed there, although by its properties it 

 is equally allied to the second. Cobalt and Nickel certainly resemble much more 

 closely the members of the first than of the second sub-series, although their atomic 

 weights place them in the second. AVith this exception, the subdivision of the series 

 which the atomic weights require does not differ from that suggested by the properties 

 of the elements. The members of this series may of course be still further subdivided 

 into groups according to their special properties, as they are in all works on Chemistry, 

 They are placed together here because the atomic weights form but one numerical 

 series. 



The isomorphism of the members of this series will be found well established with 

 the limitations before given. In order to establish the isomorphism of Cobalt and 

 Nickel with Iron, the isomorphism of one atom of Arsenic with two atoms of Sulphur 

 has been assumed. This is generally admitted ; but if it is not, no one can doubt in 

 regard to the isomorphism of these three metals, as they constantly replace each other. 

 Glucinum, Zirconium, Lanthanum, Cerium, and Thorium cannot be shown to be iso- 

 morphous with the other metals by any of their compounds, but their oxides are known 

 to replace the analogous oxides of the other metals. So also is Ruthenium known to 

 replace Rhodium. There have been doubts expressed in regard to the existence of a 

 monometric form of Zinc ; but as we have established its isomorphism with the other 

 members of the series, not only by its own crystalline form, but also by those of its 

 compounds, the fact is of no importance to the present question. 



The atomic weights of the members of this series, as determined by observation, very 

 nearly correspond with the theoretical numbers, which is the more remarkable, as the 

 limit of error in the determination of the atomic weights of the greater number, espe- 

 cially of the rarer metals, is quite wide. 



