On June 9, 1898, we had the honour of giving the Society a preliminary account of “Krypton, a new Constituent of Atmospheric Air” ; and on June 16, we brought to the notice of the Society the existence of two other gases in air, which we named “Neon” and “Metargon.” And at the meeting of the British Association at Bristol, we gave a further short account of these gases, and also announced the discovery of yet another gas, still heavier than the former, which we named “Xenon.” In a note to ‘Nature’ in October, 1898, Mr. Baly called attention to the presence of helium lines in the spectrum of neon; the D 3 line had been previously remarked in atmospheric air by Professor Kayser, of Bonn (‘Chemical News,’ August 23, 1895), and by Dr. Friedländer, of Berlin ( ibid ., December 9, 1895). Since publishing the preliminary notes referred to, we have been engaged in a study of these gases; and we may say at the outset that we have failed to isolate “Metargon.” It will be remembered that Professor Schuster, in a letter to ‘Nature’ (vol. 58, p, 199, June 30, 1898) pointed out that the spectrum of metargon presents a close analogy with that of carbon in the “Swan” spectrum. This we verified, and subjected the gas to prolonged sparking with oxygen in presence of caustic soda, attempting thus to oxidise and remove the carbon compound. The carbon spectrum was, however, still visible in a “PIiücker” tube at a pressure of several millimetres of mercury. We were inclined to ascribe this spectrum to the presence of a compound of carbon with one of the usually inert gases. But we have since failed to detect this spectrum; and we think it probable that in removing oxygen from the sparked gas some carbon was introduced with the phosphorus we then used. Remsen (‘American Chemical Journal,’ V, No. 6) has drawn attention to the fact that some specimens of yellow phosphorus contain carbon, probably in combination, and that on burning the phosphorus, carbon monoxide and dioxide are formed. After the supposed purification from carbon compounds, we must have introduced carbon monoxide into the argon of which the bulk of the sample of gas consisted, and still recognising a spectrum similar to that of carbon, we ascribed it to the presence of “metargon.” The phosphorus we at present employ does not produce such a spectrum, and we must therefore withdraw our statements regarding this supposed gas.