30 Mr. T. Bayley on the Atomic Weight and the 



circumstances that admit of comparison under equal condi- 

 tions. Thus we do not compare the highest atomicity of an 

 element with the highest atomicity of another element, or 

 their atomicities in different compounds and under widely 

 different conditions; but we compare the atomicity of the two 

 elements in those cases where they are united to the same 

 elements — that is, if they are of similar character. And we 

 generally compare them in those compounds in which the two 

 constituents are of oppositive properties (positive and nega- 

 tive) and have most completely balanced each other's affi- 

 nities. And if we take this last consideration always into 

 account, it sometimes follows that the subfamilies are found to 

 have, besides their resemblances, important structural differ- 

 ences, in which case the family relationship may be considered 

 almost to disappear. The sodium, copper, and silver group, 

 in this way considered, betray, besides the structural resem- 

 blances of their lower (proto)chlorides, oxides, &c, more im- 

 portant differences in their normal (that is, most stable and 

 evenly balanced) compounds. It may be said that, as the 

 atomic weight increases and the cycles become of greater 

 magnitude, natural families differentiate, splitting first into 

 two and then into four branches, the divergence in properties 

 each time becoming greater and the resemblances less distinct. 

 But in each family there is one branch which preserves unim- 

 paired the original main characteristics of the family. 



We may now consider what are the causes of this last phe- 

 nomenon, this maintenance of the original conditions by one 

 branch of a family ; but before doing so, we may indicate 

 some circumstances that are very probably connected with the 

 smaller differences between the members of a family or sub- 

 family of elements. We refer to the proportional position of 

 the elements in the series, or, in other words, to the intervals 

 of atomic weight at which the elements of each series occur. 

 Thus, if we represent increment of atomic weight in each 

 series by a line, we shall find that the lines representing the 

 various series are not proportionally divided, analogous ele- 

 ments occurring at unequal distances from the commencement 

 of the line, and the spaces between pairs of analogous elements 

 being disproportionate. These circumstances, it is conceivable, 

 may be connected with irregularities in the properties of the 

 most closely allied elements. 



Passing on to consider the conditions that are associated 

 with and, it may be, cause the resemblances between the 

 members of the main branch of a family, we notice that they 

 are all similarly situated in the cycle. We may, for conveni- 

 ence, distinguish the cycles, the series, and their individual 



