The magma occasionally comes into view in volcanic eruptions, and the confined gases are then partly given up into the air, but they are also partly retained in the solidifying lava, or volcanic rocks, whence they may be driven off by high temperature and subsequently analyzed. The direct gas emanations from the craters may also be gathered and analyzed. Such investigations have been carried out on a large scale by Albert Brun, Frenchman, Arthur Day, American, and his co-labourers Shepherd and Perret. Brun reached the surprising conclusion that water vapour, hitherto considered the most important of the volcanic gases, in reality was not one of them but originated in the crust of the earth. This theory, however, was completely refuted through the investigations by Day and his associates. As an example, we give the analysis (the mean of several determinations) of volcanic gases from the crater Halemaumau on the volcano Kilauea in Hawaii: _May, 1912_ _Per cent. of Volume_ Carbon dioxide 55.4 Carbon monoxide 4.3 Hydrogen 7.7 Nitrogen 29.6 Sulphurous acid 2.9 _December, 1912_ _Per cent. of Weight_ Carbon dioxide 42.9 Nitrogen 25.8 Water 27.5 Sulphurous acid 3.7 In the latter case, it has been shown that air had gained entrance to the volcanic gases, which air might have carried a quantity of water. But this quantity could not have been large judging by the amount of nitrogen present which corresponds to not quite 3 per cent. water by weight. In the former case water was not included in the analysis. At any event, a high percentage of water has frequently been observed in volcanic gases. When the gases were left in contact with water a considerable part was absorbed thereby, particularly compounds of chlorine and fluorine and also ammonia and sulphurous acid. An analysis of such water showed 10 per cent. more fluorine than sulphurous acid and two fifths as much chlorine as fluorine. The ammonia amounted only to half of one per cent. of the chlorine. None of the rare air gases were present, which indicates that the nitrogen originated entirely from the magma and not from the air. Brun has analyzed lavas from different volcanoes. The gases extracted therefrom naturally do not give as reliable information about the original atmosphere of the Earth as do the gases directly emanating from the volcanoes. As examples we quote the composition of the vapours in lavas ejected March 4, 1901, from Stromboli, and from Vesuvius in the well-known eruption of 1906. They show in percentages of volume: