56 ROSA, VINAL, AND McDANIEL: SILVER VOLTAMETER 



and then several portions of silver nitrate filtered thru the 

 pores until equilibrium between the pot and solution is obtained 

 since at the beginning the pot acts as a catalyzer, to reduce a very 

 small amount of solution. After being properly prepared there 

 is no evidence of stenolysis or other abnormality in the ordinary 

 use. 



The question of the purity of the deposited silver is an impor- 

 tant one, but the conflicting evidence in the literature does not 

 enable one to decide even whether inclusions are present or not. 

 The uniformity of the results that we have obtained in three series 

 of experiments suggests that the inclusions must be very small 

 or surprisingly constant in amount. It seems to us that inclu- 

 sions if present would be between crystals rather than in a crystal 

 perfectly formed according to the cubic system and hence greater 

 inclusions are to be expected in deposits which are less crystaline 

 because of the action of colloids. 



We have endeavored to detect the presence of AgNOs in the 

 deposits by amalgamating them with mercury then washing the 

 amalgam and testing for nitrates by phenol sulphonic acid. Our 

 blank experiments, adding AgNOs to the amalgam after precipi- 

 tating any mercury and silver salts present worked satisfactorily, 

 but we have been unable to detect any appreciable AgNOs in the 

 best deposits altho a trace was found in striated deposits. 

 If considerable filter paper was used some organic matter was 

 visible after amalgamating the deposit. We tried to determine 

 the water content of a striated deposit having an excess weight 

 over a normal deposit of 0.56 mg. by alloying the deposit to the 

 platinum crucible in which it was made over an alcohol blast, and 

 found a loss of 0.053 mg. which previous experiment showed was 

 not due to the change in the platinum crucible alone. We were 

 not able at this time to make further experiments and therefore 

 do not consider the result conclusive. 



The reactions in the voltameter have not generally been con- 

 sidered as reversible, that is, an equivalence of the loss in weight at 

 the anode and gain at the cathode, but since our previous work has 

 shown that the supposed complex reactions at the anode do not 

 probably exist we examined the possibility of reversibility. For 

 an anode we used a hea\^^ deposit contained in a large platinum 



