It is convenient to refer here to the share which Stokes and Thomson took in the physical explanation of the dark lines of the solar spectrum, and to their prediction of the possibility of determining the constitution of the stars and of terrestrial substances by what is now known as spectrum analysis. Thomson used to give the physical theory of these lines in his lectures, and say that he obtained the idea from Stokes in a conversation which they had in the garden of Pembroke at Cambridge, "some time prior to 1852" (see the quotation from his Nature article quoted above, p. 80, and the _Baltimore Lectures_, p. 101). This is confirmed by a student's note-book, of date 1854, which is now in the Natural Philosophy Department. The statements therein recorded are perfectly definite and clear, and show that at that early date the whole affair of spectrum analysis was in his hands, and only required confirmation by experiments on the reversal of the lines of terrestrial substances by an atmosphere of the substance which produced the lines, and a comparison of the positions of the bright lines of terrestrial substances with those of the dark lines of the solar spectrum. Why Thomson did not carry out all these experiments it would be difficult to say. Some of them he did make, for Professor John Ferguson, who was a student of Natural Philosophy in 1859-60, has recently told how he witnessed Thomson make the experiment of reversing the lines of sodium by passing the light from the salted flame of a spirit lamp through vapour of sodium produced by heating the metal in an iron spoon. A few days later, says Professor Ferguson, Thomson read a letter to his class announcing Bunsen and Kirchhoff's discovery.