[Sidenote:  Elementary bodies]

The greater number of the so-called ‘elementary’ bodies, now known, had been discovered before the commencement of our epoch; and it had become apparent that they were by no means equally similar or dissimilar, but that some of them, at any rate, constituted groups, the several members of which were as much like one another as they were unlike the rest.  Chlorine, iodine, bromine, and fluorine thus formed a very distinct group; sulphur and selenium another; boron and silicon another; potassium, sodium, and lithium another; and so on.  In some cases, the atomic weights of such allied bodies were nearly the same, or could be arranged in series, with like differences between the several terms.  In fact, the elements afforded indications that they were susceptible of a classification in natural groups, such as those into which animals and plants fall.

[Sidenote:  fall into different series.]

Recently this subject has been taken up afresh, with a result which may be stated roughly in the following terms:  If the sixty-five or sixty-eight recognised ‘elements’ are arranged in the order of their atomic weights—­from hydrogen, the lightest, as unity, to uranium, the heaviest, as 240—­the series does not exhibit one continuous progressive modification in the physical and chemical characters of its several terms, but breaks up into a number of sections, in each of which the several terms present analogies with the corresponding terms of the other series.

Thus the whole series does not run: 

a, b, d, e, f, g, h, i, k, &c.,

but

a, b, c, d, A, B, C, D, alpha, beta, gamma, delta, &c.;

so that it is said to express a periodic law of recurrent similarities.  Or the relation may be expressed in another way.  In each section of the series, the atomic weight is greater than in the preceding section, so that if w is the atomic weight of any element in the first segment, w+x will represent the atomic weight of any element in the next, and w+x+y the atomic weight of any element in the next, and so on.  Therefore the sections may be represented as parallel series, the corresponding terms of which have analogous properties; each successive series starting with a body the atomic weight of which is greater than that of any in the preceding series, in the following fashion: 

d       D         delta
c       C         gamma
b       B          beta
a       A         alpha
-      -----     ---------
w    w + x   w + x + y

[Sidenote:  The possibility of a primary form of matter.]