BULLETIN OF THE UNITED STATES FISH COMMISSION. 153 



The galvanometer showed a decided current in all of the experiments, 

 the pole attached to the protected plate being positive. Magnetic 

 oxide is therefore electro-negative to unprotected iron, as was to have 

 been inferred from the known electric relations of the protoxide and 

 sesquioxide of which it has been supposed to be made up. The current 

 is scarcely, if at all, stronger when two cells are connected iu series than 

 in each of them examined singly. The unprotected iron (electro-positive) 

 element wastes away, remaiuiugclean excepting for a thin, greenish, semi- 

 gelatinous film (hydrated ferrous carbonate), while the cell fluid becomes 

 turbid within twelve hours from the accumulation of sesquioxide of iron. 

 In three out of four cells small masses of sesquioxide adhered to the sur- 

 face of the protected (electro-negative) element, and were found to mark 

 spots of disintegration and removal of magnetic oxide. In one cell the 

 protected surface remains unaffected after seventy-two hours of closed 

 circuit, the fluid being densely turbid by the accumulation of sesqui- 

 oxide. 



When diluted hydrochloric acid (about 4 per cent, strength) is used 

 as an exciting fluid, the action is much more vigorous and both plates 

 are attacked, the unprotected more actively than the protected. When 

 copper is substituted in a sea- water cell for unprotected iron, the direc- 

 tion of the current is reversed, copper being electro-negative to mag- 

 netic oxide, and the magnetic oxide is rapidly decomposed, with the pro- 

 duction of sesquioxide. Iron is found (after fourteen hours) deposited 

 upon the copper surface and in solution in the filtered cell fluid. ISTo 

 copper is to be found upon the iron surface or in solution. 



It appears from these experiments that there is galvanic action be- 

 tween unprotected wrought-irou and the magnetic oxide of iron in pres- 

 ence of sea- water, at the expense of the unprotected iron. The action 

 differs only in degree from the ordinary rusting of iron, which has been 

 well described as a galvauic process from the moment that a particle of 

 sesquioxide has been formed. The begiuniug of rust is generally de- 

 termined by the presence of carbonic acid, which produces hydrated 

 ferrous carbonate, and is in turn displaced by oxygen in solution in the 

 water. From the moment of the appearance of sesquioxide of iron, a 

 compound which is electro-negative to iron, galvanic action begins, and 

 determines a further supply of oxygen by electrolysis of the water. So 

 in the cells here referred to hydrated ferrous carbonate is found upon 

 the positive surface, and sesquioxide of iron free in the liquid or adher- 

 ing to the negative surface. Internal currents result from the inter- 

 action between iron, its carbonate, and its oxides, and diminish the re- 

 sultant current strength as measured by the galvanometer. 



The experiments still indicate that the covering of magnetic oxide is 

 not thick enough or not complete, since most of the specimens have 

 broken down more or less in sea-water. If thoroughly eoated there 

 seems to be no reason to fear damage to protected iron near unprotected 

 iron in sea-water, the tendency of galvanic action in that case being 



