216 
in their atomic weights and physical properties admit of 
explanation by the assumption of their being allotropic 
varieties of each other, then gold, palladium, and iron, may 
stand at the head of their respective groups, and determine 
the species to which the varieties belong. It is no objection 
to the theory of the members of the respective groups being 
varieties of each other, that they cannot by any known 
power of analysis be resolved into their primaries, as the 
same objection would apply to the natural varieties of 
organic species determined by naturalists. 
We have seen that the quantivalence of most of the mem- 
bers of the preceding groups H%, H 6n, is in some way 
correlated or dependent on the construction of the typical 
molecules at the head of each series ; but in the series W7n, 
the only element which is known to be septivalent is man- 
ganese; but the relation which this metal has to the iron 
group; and bearing in mind that the determination of the 
highest quantivalence of elements is limited by the know- 
ledge of chemists at particular times, and is only arrived at 
after much research, the septivalency of manganese indicate 
a much higher quantivalence for the other members of the 
series Win, than has up to this time been accorded to them. 
I have hesitated to introduce hypothetical elements alter- 
nating with the iron, palladium, and platinum groups, as the 
regular sequence of elementary forms is broken by varieties, 
and from the density of the typical molecule H7, it may be 
that the members of this series are limited to those shown 
in the table. The density of the typical molecule may 
also explain the absence of members alternating with Cr, Mo, 
and W, and I have therefore only introduced one hypothe- 
tical element in this series, the analogue of Cr with the 
atomic weight = 144. 
Considering how nearly the numbers representing the 
molcular constitution, and atomic weights of the members 
in homologous positions in the higher groups approximate, 
