W. Gibbs on the theory of Atomicities. — 411 
With these preliminaries I shall consider in order a few cases 
of what is termed molecular union. 
Since sodium may be regarded as feebly triatomic it follows 
that in chlorid of sodium, Na™Cl, there are still two units of 
affinity unsaturated, the intensity of the affinity being very small. 
Hence the salt, Na™Cl+(20H,)”, in which the sodium may be 
regarded as saturated, exists only at low temperatures. In sodic 
sulphate we may for convenience take sulphur as the determin- 
ant, in which case the salt will have the formula, S*" 1 af a0) 
considered as anhydrous. But the crystallized salt contains ten 
2 . . 
atoms of water and may be written 5% < (Na6)', to exhibit the 
2 10 
manner in which the water of crystallization completes the sat- 
uration of the hexatomic sulphur. 
Trisodic ortho-phosphate contains twelve atoms of water, 
and if we take phosphorus as the determinant may be written 
It 
8 
pe | Bae), 
. ; (OH)’;. 
phosphorus being heptatomic. ee 
These formulas will suffice to exhibit the application of the 
theory to cases in which water of crystallization is present. To 
avoid complication in expression it would perhaps be more con- 
Venient to indicate the subordinate atomicity by Roman nume- 
rals placed beneath the symbol of the determinant and to con- 
nect the water of crystallization in the usual manner by the 
Sign +. Thus sodic sulphate may be written 
iv aot 
5 } fag), +1024 
~ or K (H™F,) 
which may be compared directly with hydrate of potash 
= Da or K(H""@). 
The so-called double fluorids of tin, zirconium and silicon may 
readily be reduced to the following types in which in each case 
the tetratomic element is taken as the determinant. 
Ss ee sn (KF), 
