420 PROCEEDINGS OF THE AMERICAN ACADEMY. 



manifest upon heating the peculiar contraction called " sinterino-," ig 

 evidence of the existence of molecular instability and possibility of re- 

 arrangement. The hypothesis of 2">artial dissociation at high tempera- 

 tures is adequate to explain this. 



It is obvious that, if the oxide is really dissociated, it should be capa- 

 ble of conducting electricity, because of the presence of traces of metal. 

 In order to test this point, small platinum electrodes of 0.25 square 

 centimeter in area and separated from one another by about 0.5 cm. were 

 packed in pure specimens of the several oxides. Each porcelain cruci- 

 ble containing an oxide was gradually heated, after having placed the 

 two electrodes in a circuit with a delicate amperemeter and four accu- 

 mulators. When cold, the resistance of each of the oxides was at least 

 200,000 ohms. Upon heating, the cupric oxide began sensibly to con- 

 duct below a red heat, and at about 650° offered only about as much 

 hindrance to the current as was afforded by 550 ohms, remaining con- 

 stant at that point. On the other hand, zincic oxide fell in resistance 

 only to about 15,000 ohms at 650°, and required fully 900° to bring 

 it as low as 1,000 ohms ; while the highest heat of the blast lamp could 

 not cause the masrnesic oxide to transmit as much as the twentieth 

 of a milliampere, the least current capable of being observed upon the 

 amperemeter. It is needless to say that any possibility of the presence 

 of a reducing atmosphere was excluded, and the fact that each oxide 

 returned almost if not quite to its non-conducting state after cooling 

 is sufficient proof that the heat aloue was the cause of the phemomena 

 observed. These interesting results are fully in accord with the hy- 

 pothesis that cupric oxide is slightly dissociated at a I'ed heat, that zincic 

 oxide is less dissociated, and that magnesic oxide, retaining its occluded 

 oxygen with great obstinacy, is scarcely dissociated at all. Magnesic 

 oxide, being one of the stablest of all compounds at high temperatures, 

 could not be expected to show any appreciable dissociation or conduc- 

 tivity ; and the wonder is, that the occluded gas is able to escape even as 

 fast as it does. 



This assumed partial thermal dissociation of solids must not be con- 

 founded with electrolytic dissociation (ionization) ; for the energy which 

 determines the separation of the components in the present case is heat, 

 and not electricity. The conductivity is not electrolytic conductivity, 

 but only the metallic conductivity of the metal set free by a purely 

 thermal dissociation. In order to prove this, a current of twenty milli- 

 amperes was transmitted through cupric oxide for three hours. This 

 quantity of electricity would have carried nearly seventy milligrams of 



