34 Mr. T. Bayley on the Atomic Weights and the 



the elements belonging to the odd series are always diamag- 

 netic." This supposition is evidently true in the first four 

 cycles, as far as the present state of knowledge enables us to 

 verify it. Thus the first cycle is paramagnetic, and the second 

 diamagnetic. The first half of the third cycle is paramagnetic 

 and the second half diamagnetic; and this is the situation in 

 the fourth cycle also. But if the fith cycle contains four periods 

 of seven elements, excluding platinum, iridium, and osmium, 

 and we know the first four cycles are strictly septenary, then 

 the series containing gold, mercury, thallium, lead, and bis- 

 muth is an even series; and yet these are diamagnetic elements. 

 This and other considerations lead to the modification of the 

 law expressing the magnetic relations. The first cycle (a 

 simple cycle of seven elements) is paramagnetic ; the second 

 cycle, also simple, is diamagnetic ; after this, in the complex 

 cycles, the first half is paramagnetic and the second diamag- 

 netic. The fifth cycle is incompletely known at present ; but 

 if subsequent research shows the first half to be paramagnetic 

 and the second half diamagnetic, it will justify the system 

 of semicycles at the expense of the system of odd and even 

 elements of Mendelejeff. 



The property possessed by certain metals of giving coloured 

 solutions when associated as bases with colourless acids is 

 strictly periodic; and the law may be stated as follows: — The 

 metals in the first and second cycles form no coloured solu- 

 tions; in succeeding cycles the metals occupying the position 

 of lowest atomic volume and the elements immediately suc- 

 ceeding them form coloured solutions. Thus in the third 

 cycle the metals from titanium to copper, in the fourth cycle 

 the metals from niobium to palladium, and in the fifth cycle 

 the platinum group gold and tungsten have this property. 

 It is interesting to observe how the power of forming coloured 

 solutions ceases abruptly in the same region of each complex 

 cycle — in the third cycle after copper, in the fourth after palla- 

 dium, and in the fifth after gold. To be consistent, silver 

 should form coloured solutions ; but its solutions are not 

 coloured, and it is not itself a coloured metal. Possibly its 

 solutions are colourless because they correspond to an oxide 

 of the form Pi' 2 0, just as the salts of copper corresponding to 

 Cu 2 are generally, but not always, colourless. If this is so, 

 salts of silver corresponding to AgO, whenever they are dis- 

 covered, may have coloured solutions. 



There is one other circumstance with regard to the position 

 of lowest atomic volume worthy of notice; it is the follow 

 ing: — The increments of atomic weight Avhich, starting from 

 hydrogen, successively give the point where the atomic volume 

 is a minimum, are members of the geometric series 



