348 PROCEEDINGS OF THE AMERICAN ACADEMY. 



IV. Intermittent Iron Arc in Oxygen and Hydrogen. 



In an atmosphere of oxygen, this persisting luminosity is very 

 much increased ; while, when the hood of the arc is filled with hydrogen, 

 the luminosity is very much decreased. In hydrogen, the light is also 

 very red and yields the H„ and B^ lines (C and F of Fraunhofer) as 

 already observed by Liveing and Dewar.* 



After the arc has been working in oxygen for a few minutes the whole 

 interior of the hood is lined with a light brown deposit, which is evidently 

 iron oxide. 



In hydrogen, on the contrary, the iron electrodes in the parts that 

 become hot present a bright metallic surface, which shows the reducing 

 effect of the hydrogen on the iron oxide ; but there is no evidence of any 

 chemical action between the iron and the hydrogen. 



So far as the brilliant luminosity of the iron arc in oxygen is concerned, 

 it is evident that it might be due either to the oxidation of the iron or to 

 the increased heat produced by this oxidation. 



In the case of hydrogen, we should say that chemi-luminescence were 

 ruled out were it not for the fact that in the hydrogen flame, which burns 

 as it escapes from the chamber surrounding the arc, there is always a 

 small white core, a flame in color not unlike that of ordinary illuminating 

 gas or acetylene. 



Since the hood was thoroughly cleansed from all forms of oil and 

 grease, and since the bearings were run dry in clean asbestos packing, the 

 only source of carbon would appear to be that which we know to be 

 ■ present as an impurity in the iron electrodes. 



That this carbon impurity is the true source of the white flame is made 

 highly probable by the fact that " chemically pure " zinc and magnesium 

 poles gives no trace of it, the carbon in these being very much less than 

 in the impure iron employed. 



V. Intermittent Arc of Pure Zinc and Magnesium 



in Hydrogen. 



To realize the conditions indicated in this caption may be con- 

 sidered the chief object of this experiment. For with a large current of 

 fresh, dry, and pure hydrogen sweeping out the hood of the arc, we may 

 feel fairly certain that oxidation is ruled out, except at the very start. 

 Any of the original air in the hood still remaining, its oxygen would be 



* Liveing and Dewar, Proc. Roy. Soc, Vol. XXXV p. 74 ; also Vol. XXX. p. 152. 



