The quickness with which a pendulum oscillates is less or greater according to its length, a long one oscillating slowly, and a short one quickly. The vibrations of a string or pipe are also slow or quick, and the note which it yields is low or high, according as it is long or short.

Similarly, according to Lecoq de Boisbaudran, the rate of vibrations of molecules, and the wave-lengths of the light which they emit, are determined by their weight. When the molecular weight is high, the vibrations of the molecules are slow, and the light which they emit has long wave-lengths, and is situated towards the red end of the spectrum. When the weight is low the vibration of the molecules is rapid; and the light they emit lies towards the violet end of the spectrum.

In the same family of elements the mean length of the wave of light which they emit is a function of their atomic weight, so that for bodies of the same chemical type the general form of the spectrum persists, but is gradually modified by the mass of the molecules. As the atomic weight diminishes, the spectrum will tend to ascend towards the violet, and as it increases the spectrum will tend to descend towards the red.

Until recently, our observations on the spectra of bodies were limited to the visible spectrum, but the application of photography now enables us to extend our observations both above and below the visible spectrum, and to ascertain the presence of definite spectra in the ultra-red and ultra-violet, when nothing of the sort is visible to the eye. In most musical sounds besides the fundamental note we have a number of harmonics having a much greater rapidity of vibration than it. Similarly, in the spectrum there appear harmonics as well as the fundamental spectral lines, and so instead of one line or band there may be a number. According to the author already quoted, the corresponding harmonics in a series of analogous spectra have mean wave-lengths which increase in proportion to the weight of the molecules.

It might appear, therefore, that a relation might be observed between the spectroscopic characters and physiological action of an element, and this idea was propounded by Papillon. His idea was, however, to a great extent based on the experiments of Rabuteau referred to later on, and just as no definite relation can be at present traced between the atomic weight and the toxic action of a metal, so no definite relation can be observed between its spectroscopic characters and its physiological action.

Further consideration, however, will show us that this is not at all to be wondered at, for in physiological experiments we are not working with the same molecules which yield the spectrum.

In spectrum-analysis, when line spectra are in question, according to one view we are in presence of phenomena produced by the chemical atom, whereas this atom exists only molecularly combined at lower temperatures. According to another view, that put forward by Lockyer, we are in presence of phenomena produced by a series - possibly a long series - of simplifications, brought about by the temperature employed, and this simplification can begin at very low temperatures, and is indeed indicated by Dalton's law of multiple proportions.

Such molecular simplifications and differences are represented by ozone and oxygen, ordinary and amorphous phosphorus, the various forms of sulphur and so on, and it is therefore at this lower range of temperature - where the phenomena are to be studied by absorption, and not by radiation - that we must look for connections between molecular structure and physiological action if any such connection exists.1

Some of the absorption bands which occur in the spectra of bodies at ordinary temperatures may be in the visible spectrum, like those observed in alcoholic and aromatic substances;2 but others may be quite invisible, and only recognisable by the aid of photography in the ultra-red or ultra-violet.3