248 THE NUMERICAL RELATION 



tions are possible in the other systems, and therefore that we cannot be expected to 

 establish isomorphism in any case except between nearly consecutive members. 



The atomic weights of the members of this series, with the exception of Phosphorus, 

 do not present any important deviations from the theoretical numbers, taking into ac- 

 count always, of course, the amount of possible error. The deviation in the case of 

 Phosphorus has already been noticed. Oxygen, it must be admitted, is not connected 

 with the series from any similarity of properties, though the Phosphides, Arsenides, and 

 Antimonides present certain analogies with the Oxides. As has already been said. Ox- 

 ygen was placed at the head of this, as well as of the last two series, because its atomic 

 weight seemed to be the nucleus of all three. 



The Five Series is the shortest of all, consisting of only three members. Carbon, Boron, 

 and Silicon. Of these, the last two are as closely allied as are any two members of the 

 other series, Silicon having precisely the properties we should expect in a homologue 

 of Boron, which was lower in the series ; and the same is also true of their compounds. 

 The analogies, however, between these two elements and Carbon are by no means so 

 close, for not only Carbon cannot be proved to be isomorphous with them, but it 

 does not form similar compounds. Carbonic Acid, it is true, presents some points of re- 

 semblance to Boracic and Silicic Acids ; like them it unites in a large variety of propor- 

 tions with bases, its alkaline salts give a basic reaction, &c. ; but according to the gen- 

 erally received opinion, its sjTnbol is CO^, while those of Boron and Silicon are BO3 and 

 SiOj. In its uncombrned state, however. Carbon resembles Boron and Silicon, not only 

 in its outward properties, but also in its action before the blowpipe. Two of the allo- 

 tropic states of Carbon, Graphite and Charcoal, are probably repeated in Boron, and 

 are known to be in Silicon. The principle of exclusion would also seem to place Car- 

 bon in this series, for it certainly presents no analogies with the members of any other. 

 The correspondence of the atomic weights of the members of this series to the law is 

 remarkably close. 



The Four Series is by far the largest of all, including the greater number of what are 

 generally known as the heavy metals. The members of the series resemble each other 

 in the following respects. First, they are isomorphous ; for although each member can- 

 not be dii-ectly proved to be isomorphous with every other, yet isomorphism can be 

 established between consecutive members, which, as has before been said, is all that can 

 be expected. Second, the members of this series all form, by uniting with Oxygen, 

 either Protoxides or Sesquioxides, or both, which, as a general rule, are strong bases. 

 Third, these Oxides are either insoluble, or nearly insoluble, in water. And finally, the 

 elements of the series have all those physical properties which are known as metallic 

 properties. 



