Three times it seemed on the verge of being anticipated. The experiment, which in Kirchhoff's hands proved decisive, of passing sunlight through glowing vapours and examining the superposed spectra, was performed by Professor W. A. Miller of King's College in 1845.[395] Nay, more, it was performed with express reference to the question, then already (as has been noted) in debate, of the possible production of Fraunhofer's lines by absorption in a solar atmosphere. Yet it led to nothing. Again, at Paris in 1849, with a view to testing the asserted coincidence between the solar D-line and the bright yellow beam in the spectrum of the electric arc (really due to the unsuspected presence of sodium), Léon Foucault threw a ray of sunshine across the arc and observed its spectrum.[396] He was surprised to see that the D-line was rendered more intensely dark by the combination of lights. To assure himself still further, he substituted a reflected image of one of the white-hot carbon-points for the sunbeam, with an identical result. _The same ray was missing._ It needed but another step to have generalised this result, and thus laid hold of a natural truth of the highest importance; but that step was not taken. Foucault, keen and brilliant though he was, rested satisfied with the information that the _voltaic arc_ had the power of stopping the kind of light emitted by it; he asked no further question, and was consequently the bearer of no further intelligence on the subject. The truth conveyed by this remarkable experiment was, however, divined by one eminent man. Professor Stokes of Cambridge stated to Sir William Thomson (now Lord Kelvin), shortly after it had been made, his conviction that an absorbing atmosphere of sodium surrounded the sun. And so forcibly was his hearer impressed with the weight of the argument based upon the absolute agreement of the D-line in the solar spectrum with the yellow ray of burning sodium (then freshly certified by W. H. Miller), combined with Foucault's "reversal" of that ray, that he regularly inculcated, in his public lectures on natural philosophy at Glasgow, five or six years before Kirchhoff's discovery, not only the _fact_ of the presence of sodium in the solar neighbourhood, but also the _principle_ of the study of solar and stellar chemistry in the spectra of flames.[397] Yet it does not appear to have occurred to either of these two distinguished professors--themselves among the foremost of their time in the successful search for new truths--to verify practically a sagacious conjecture in which was contained the possibility of a scientific revolution. It is just to add, that Kirchhoff was unacquainted, when he undertook his investigation, either with the experiment of Foucault or the speculation of Stokes.