Chemical and Physical Properties of Elements. 37 



cyclic position, especially among the higher atomic weights. 

 Now, as the cycle increases, the same progression of atomic 

 weight produces less and less change in cyclic position, and 

 consequently less in properties. Comparing the best-known 

 members of the fifth cycle with the best-known members of 

 the third and fourth, we remark that the progression of atomic 

 weight from gold to bismuth is only about 1 1 , while the pro- 

 gression from silver (107-66) to antimony (120) is about 12, 

 and from copper to arsenic 11*9; and this, although the semi- 

 cyclic progression from caesium to gold, viz. 66, is three 

 times the semicyclic progression from rubidium to silver, 

 viz. 22*4. Therefore, while the dimensions of the fifth cycle 

 are three times those of the fourth, the mean progression of 

 atomic weight between adjacent atoms is about the same ; 

 the obvious inference is the presence of three times as many 

 atoms in the fifth cycle : and upon these facts an argu- 

 ment may be founded to prove the fifth cycle to contain six 

 primary series of seven elements. This supposition receives 

 support from the closeness in atomic weight between cerium 

 and lanthanum and between tantalum and tungsten, and also 

 from the number of newly discovered elements, many of them 

 earth metals, waiting to be classified*. 



On the other hand, this equality in atomic progression does 

 not hold good at the extremities of the cycles ; and in these 

 positions the atomic progressions are greatest. There is 

 between potassium and calcium a difference of 1, between 

 rubidium and strontium a difference of 2, while the progres- 

 sion from caesium to barium is 4. Again, small progressions 

 of atomic weight between adjacent elements in a large cycle 

 might, it is conceivable, be insufficient to bring about transi- 

 tion from one full atomic type to another, and might result in 

 the formation of twin or triplet elements not sufficiently dif- 

 ferent to have separate atom-analogues in the lower cycles. 

 Cerium, lanthanum, and perhaps didymium may thus have 

 common atom-analogues in yttrium, scandium, and alumi- 

 nium, and uranium and tungsten in molybdenum and chro- 

 mium. In this way the fifth cycle may have four or even only 

 two internal periodic variations, and yet have more than thirty- 

 one elements. Again, tungsten may be the atom-analogue of 

 manganese and of the unknown element IY. a 7. However, 

 these are questions too obscure to be answered in the present 

 state of chemistry, and await the light of further investigation; 

 but this seems clear, that we must expect to find in the fifth 

 cycle developments not to be found in the lower ones. 

 * Some of these may form part of a sixth cycle. 



