under extremely low temperature. And it may be supposed that during this period other gases may exist associated with these, which, however, unlike them, are unable to sustain the successively higher and higher temperatures which the nebula acquires in its process of condensation and organization into a system, and at certain stages of this process are dissociated and resolved into aggregates of a different constitution, suited to these temperatures. Some of these latter new aggregates would naturally assume the liquid and solid forms at temperatures still high as compared with those to which we are accustomed, and constitute in the cooled-off crust of the planets the various metals and metalloids. In this manner we should have no difficulty in accounting for the existence of all the elements found on the earth, even if it were positively known that only the lighter gases were present in the parent nebulæ.
The recognized elementary substances, presenting so many different qualities, vary greatly in their so-called "atomic weights." This means simply that their molecules vary greatly in mass. The hydrogen molecule is the least known, and is therefore taken as the standard. Compared with this as unity, we find that the molecule of oxygen contains 16 times as much matter, that of carbon contains 12, that of nitrogen 14, and that of chlorine 3512 times as much. But these, instead of representing large equivalents, are, when compared with most of the metals, very small. One molecule of mercury contains two hundred times as much matter as one of hydrogen. The atomic weight of gold is 197, of platinum 197⋅4, of lead 207, and of bismuth 208; while the thorium equivalent, which was quadrupled in the new system, is now put at 231⋅4, being the largest of all the elementary units. Whether hydrogen, carbon, nitrogen, oxygen, or any of the other abundant elements having small molecules have entered into the composition of these heavy substances, is a legitimate question. The fact that these molecules are stable, whether combined or uncombined, is favorable to this view, although there may exist, as component units of the molecules of the metals, many equally stable aggregates which no human power can dissociate from their present combinations. But, if known elements were employed as components of other known elements having larger molecules, the very fact that they are elements, i. e., that we are unable to decompose them, would render it impossible to know that such was the case. Think how many hydrogen, nitrogen, or oxygen molecules might enter into the system that constitutes the unit of bismuth or of gold!
Now, it is a remarkable fact that those elements which have very high condensing points, i. e., which assume the liquid (or solid) form at very high temperatures, generally have large combining numbers, that is, large molecules; while those having low condensing points and which are gaseous at ordinary temperatures, as a rule have small combining numbers, or small molecules. To this, carbon on the one hand,
