410 W. Gibbs on the theory of Atomicities. 
primary or inherent atomicities are represented by odd, and 
those whose atomicities are represented by even numbers. __ 
2. The atomicities are variable and no maximum or limit 
ean be assigned in any case. Any element of the first class may 
therefore be 1, 3,5, 7, &c., atomic, and any element of the 
second class 2, 4, 6, 8, &c., atomic. : 
8 the combinations of an element with itself or with 
another element a certain number of units of atomicity disap- 
pear. The atomicity of the resulting compound will be termed 
secondary. 
For the sake of convenience I shall give the name of deéer- 
minant to the dominant element in any compound, the other 
elements being termed sub-determinants or subjects, while the red 
pound itself is termed the resultant, Thus in ammonia, N< H, 
nitrogen is the determinant, and hydrogen the subject, while 
€u 
in nitrid of copper, N, { €u, copper is the sub-determinant. In 
€ 
borid of nitrogen, B™'N"™", either element may be taken as the de- 
terminant. According to this view ammonia may be regarded 
as the product of a single action, that, namely, of nitrogen upon 
hydrogen, while in nitrid of copper there are two actions, the 
duplication of the atom of nitrogen being due to the diatomic 
character of the copper. ‘ 
If now we consider first the case of water, OH,, it is easy to 
see that if oxygen be regarded as tetratomic or at least as hav- 
ing an even atomicity higher than 2, water will itself be weakly 
diatomic, whatever may be the number of atoms of this body 1” 
a given compound. For in OH, it is clear that only two ~— 
f affinity are saturated. Morever, in any number of atoms 0 
pease i units will vanish for every atom with the exception 
of the 
Le OMe ee. een ee eee 
rst and last, in each of which one unit only will disap” 
pear. Graphically this may be represented as follows: 
2 OOOO —-0- 0+ 
From this it appears that the atomicity of the water of ote 
tallization in a salt may always be considered as 2, whatever ™ 
the number of atoms. In like manner it is easily seen that eer q 
mona, NH,, is diatomic, and that any number of associ# eo 
atoms of ammonia will still, as a whole, represent only a simg® 
_ diatomic atom. oS aca 
