204 



SCIENCE. 



[N. S. Vol. XXIV. No. 607. 



has a density of 2.654 (25°). Neither 

 CaaSisOg, Ca^SioOio nor 3CaOSi02 can ex- 

 ist in the two-component system. 



The Effect of Acetone on the Transference 

 Numbers of Sodium and Potassium 

 Chlorides: H. F. Lewis. 

 It has been found that in general a 

 change of solvent has little or no effect on 

 the transference number of a salt, but no 

 experiments have been published in which 

 acetone was used. In the present investi- 

 gation the apparatus was a large inverted 

 U tube, the legs of which dipped into 

 large test tubes; by means of a small tube 

 blown into the top or bend of the U tube 

 it was possible to withdraw a sample of 

 the middle portion for analysis. Although 

 the apparatus is not at all adapted to very 

 accurate work, it was shown by blank meas- 

 urements and tests with indicators that the 

 method was sufficiently good to guarantee 

 that the large differences found are not due 

 to errors. A silver voltameter was used 

 to measure the total decomposition. The 

 cathode portions were analyzed by titrating 

 with deci-normal silver nitrate. The 

 middle portion changed in almost all eases 

 less than one per cent. 



NaCl 



KCl 



In the above table of results the first 

 column gives the parts of acetone in one of 

 solution; the second column gives the 

 transference number of the chlorine, and 

 the last column gives the number of inde- 

 pendent experiments on which the result is 

 based. The experiments were carried out 



at room temperature ; and all the solutions 

 were approximately deci-normal with re- 

 spect to the salt. 



The Electrode Capacity of Iron and its 



Bearing on Passivity: C. McC. Gordon 



and Friend E. Clark. 



The oldest and most commonly accepted 

 explanation of the passivity of iron carries 

 with it the idea that it is due to a very 

 thin oxide sheet. In later years several 

 other hypotheses have been suggested, 

 which, while perhaps better fitted to ex- 

 plain many of the existing conditions, have 

 not entirely displaced the oxide theory. In 

 ease we assume the oxide theory we might 

 expect the so-called passive iron to act as 

 an electric condenser; the conducting iron 

 and the conducting solution being sepa- 

 rated by insulating oxide, similar to films 

 on aluminum anodes. Measurements of 

 this capacity — comparing it on the Wheat- 

 stone bridge with a metallic condenser of 

 known capacity— have been made for iron 

 electrodes in various solutions, with the 

 following results: 



(1) Passive iron (electrodes) acts as if 

 covered with a sheet of oxide, as is evi- 

 denced by the following facts: (a) The 

 electrodes have a capacity easily measured 

 while active electrodes (that is, iron in 

 dilute nitric acid or freshly deposited elec^ 

 trolytic iron) show no signs of such capaci- 

 ties. (&) The capacity values are about 

 of the same order of magnitude as those 

 of aluminum, where we certainly know 

 there is such capacity. (c) Electrodes 

 oxidized in a Bunsen burner flame give a 

 similar capacity. The oxidized sheet, so 

 formed, appears to be five or six times as 

 thick as that of the electrodes made passive 

 by the concentrated nitric acid. It is no- 

 tably different, too, in that it has a small 

 resistance. It acts like a leaky condenser, or 

 a condenser in parallel with a resistance. 

 (2) Iron transferred from the air to neu- 



