Washington — Submarine Eruptions of 1831 and 1891. 149 



that C0 2 can be reduced to CO by the action of hydrogen at 

 elevated temperatures. Gautier also showed that the reaction 

 is reversible, at a white heat being : 



C0 2 + H 2 = CO + H 2 0, 



and at temperatures between 1200° and 1250 ° 



CO + H 2 = C0 2 + H 2 . 



As Clarke remarks : "When water emitted by heated rocks 

 (or that of the sea, H. S. W.) mingles with carbon dioxide, 

 within the vent of a volcano, both reactions take place, and 

 mixed gases, which sometimes contain a trace of formic acid, 

 are generated. This mixture is a powerful reducing agent, 

 which acts upon the iron silicates in an opposite direction to that 

 of the oxidizing vapor of water. Either oxidation or reduction 

 is therefore possible, according to the preponderance of one 

 constituent or another among the volcanic gases." 



We may therefore ascribe the difference in the iron oxide 

 ratios in the scoriaceous lavas to the fact that, in one case they 

 were erupted subaerially and in the other subaqueously. In 

 the former large quantities of heated air would have acted 

 on the highly vesicular masses during their ejection, and would 

 have materially aided the steam present in oxidizing the ferrous 

 iron of the magma. In the other, free atmospheric oxygen 

 was wanting and, as we know that carbon dioxide was emitted 

 in large amount at Graham Island, and presumably also at 

 Foerstner Volcano, we may suppose this and the steam to have 

 reacted with the formation of carbon monoxide and hydrogen, 

 which would have had a powerful reducing effect on any ferric 

 iron present in the magma. In the case of the compact and 

 scarcely vesicular lavas, such as those of Catalonia, which were 

 ejected in the form of massive flows, there would have been 

 comparatively little opportunity for either an oxidizing or a 

 reducing action, and we should expect to And, as we do, ratios 

 of ferrous to ferric oxide intermediate between those of the 

 scorias, and presumably more nearly like that of the original 

 magma. 



The reducing conditions at the submarine eruptions must 

 also have been very materially increased by the hydrogen sul- 

 phide and sulphur dioxide, which we know to have been 

 present. In this connection it is of interest to recall that John 

 Davy found that the waters of the two small craters of Graham 

 Island, on August 5, contained hyposulphite of lime and 

 magnesia, in addition to sulphates, but no salts of the alkalis. 

 Although the condition of chemical analysis at the time does 

 not give this statement much weight, yet it may be significant. 

 Abich remarks on the absence of hydrochloric acid during the 



Am. Jour. Sci.— Fourth Series, Vol. XXVII, No. 158.— February, 1909. 

 11 



