volcanic ejections. After the mighty eruption of Krakatoa, for example, I sought in vain for these elements in the cinders cast out, which probably came from great depths. The supposed discovery of a new element in the ancient lavas of Vesuvius has been found to be erroneous.
Elementary bodies seem to multiply as we approach the surface of the globe. Two hypotheses suggest themselves in explanation of the fact: that of displacements of cosmical matter, and that of the new formation of elements on the surface.
The displacements of cosmic materials are incessant; falls of meteorites furnish a particularly striking example of them, but it is probable that as to quantity the cosmical dusts are of more importance. Yet neither the meteorites found at various points nor the dust collected by Nordenskiold in the ice fields of the polar regions, the extra-terrestrial origin of which can not be doubted, contain the rare elements of the earth. The hypothesis of an increase by accretions from without appears to lack foundation.
The new formation of elementary bodies seems to be still less probable; at most it might be explained by the possibility, often indicated but never established, of a new reduction of bodies heretofore supposed to be simple. Spectrum analysis, it is true, reveals to us transformations which are gradually going on in the matter of the fixed stars, but they are only of known substances becoming converted into other substances equally known. Moreover, the conditions of temperature and aggregation of the fixed stars and those of the earth can not be compared.
It is evident that the increase of simple bodies in the outer strata of the earth is only apparent. It should be recognized, besides, that science has made great progress, and this progress can not be without influence on the discovery of new substances. The first electrolytic decompositions accomplished by Davy with an inferior voltaic pile made known at the beginning of this century the existence of metallic radicles in the salts and the earth of which there had not been before the slightest suspicion; while Moissan, by the employment of the powerful currents now available, has been able to disengage fluorine—hitherto almost unknown—from its combinations. Spectrum analysis has cast light on a whole series of elements of characteristic spectra. The presence of one of these elements, helium, had been demonstrated in the sun before it was known that it likewise entered into the composition of our globe. The conclusions drawn by D. Mendeleef from the periodical law have also led to the discovery of several elements the existence of which was indicated by theory before the chemist had isolated them. I mention, first, scandium, discovered in 1879 by Nilson in exonite,
