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



has analogies with magnesium, and the newly-discovered gal- 

 lium with aluminium ; after gallium the metallic tendency 

 grows weaker and the atomicity higher, in arsenic; and the 

 negative monad type culminates once more in bromine. 

 Again there is a break, followed by a metal of the alkalies — 

 rubidium. We thus recognize the completion of the full 

 period; but it is double, and consists (if we omit the interme- 

 diate elements iron, cobalt, and nickel) of two series of seven. 

 The increment from lithium to sodium and from sodium to 

 potassium was sixteen ; the increment from potassium to 

 rubidium is approximately three times sixteen — viz. 46*4. 

 From rubidium to the next alkali metal (caesium) the incre- 

 ment is 47*5 — almost the same as 46*4 ; and the intermediate 

 elements form a cycle parallel in a remarkable manner with 

 the cycle from potassium to rubidium. Each of these twin 

 cycles is double, consisting of two series of seven united by 

 an intermediate group, consisting in the second case of palla- 

 dium, rhodium, and ruthenium. The element silver, at the 

 head of the second series of the fourth cycle, has long been 

 classified with copper, and cadmium with zinc; and the resem- 

 blances between antimony and arsenic, between tellurium and 

 selenium, and between bromine and iodine, are among the best- 

 known facts in chemistry. 



There is no alkali metal known of higher atomic weight 

 than caesium ; but there is strong evidence that, just as the 

 third and fourth cycles succeed the first and second, and cor- 

 respond to increment of atomic weight three times as great, 

 so the third and fourth cycles are succeeded at any rate by a 

 fifth, with increment of atomic weight again three times as 

 great. The fifth cycle is incomplete, as indeed are the third 

 and fourth in a few members; but it seems to consist of four 

 series of seven, united two to two by the metals osmium, iri- 

 dium, and platinum, which closely resemble palladium, rhodium, 

 and ruthenium. 



It will be seen that, while the cyclic change retains its cha- 

 racteristics unimpaired from the lowest to the highest known 

 atomic weight, the series, which in its early form is cyclic, loses 

 its prominent features as the atomic weight progresses. But 

 besides the natural classification into cycles and series, the 

 elements are capable of another mode of grouping. Certain 

 elements have long been seen to have strong points of resem- 

 blance and analogy, and have therefore been grouped into 

 natural families. The alkali metals, the metals of the alkaline 

 earths, and the halogens, are the best examples of family 

 groups. A moment's consideration of these typical examples 

 will show that families consist of elements which have corre- 



