72 ROYAL SOCIETY OF CANADA 



ered from different points on the neck. AVeighings were thus made for 

 about every centimetre, and these distances divided into tenths. All 

 points were then tested and corrections applied if necessary. 



To find out how accuratel}^ the volume of a constituent of a mixture 

 could be measured, the following measured volumes of water which, 

 according to pipette readings, should be equal, were weighed. The vol 

 umes given by weighing were 30*550 c.c, 30'560 c.c, 30-564 c.c, 

 30-5-40 c.c; mean=30'553. The probable error would seem to be less than 

 0«05 per cent. 



Determining the Conductivity. 



The Kohlrausch method with the telephone and alternating cur^ 

 rent was used. The measuring apparatus consisted of four resistance 

 coils, and a German silver bridge- wire, about three metres long, wound 

 on a marble drum. The wire was divided into 1,000 parts and had a 

 resistance of about 114 ohms. It Avas calibrated by the method of 

 Strouhal and Barus.' 



The resistance coils were marked 1, 10, 100 and 1,000 ohms. As I used 

 only one coil (that of 1,000 ohms), and as it was not necessary to express 

 the conductivities in absolute measure, 1 did not need to know the 

 relative accuracy of the coils or the absolute value of the one used. 



Two electrolytic cells were used, one for strong solutions, the other 

 for solutions more dilute. They were of the U-form, shown by Ostwald 

 iji his Physico-Chemical Measurements, p. 226, fig. 1*78. 



The electrodes were of stout platinum foil not easily bent, circular in 

 form and about 3*5 cm. in diameter. Care was taken always to have the- 

 electrodesas nearly in the same position in the electrolytic cell as possible» 

 No change in conductivity could be observed for small differences in 

 position, such as could be detected and avoided. 



The induction coil was small and had a very rapid vibrator. It was 

 kept in a box stuffed with cotton wool, that the noise might not interfere 

 with the determination of the sound minimum in the telephone. A Le- 

 clanchC' cell Avas found most convenient for Avorking the coil. With this 

 arrangement the minimum point on the bridge Avire could be determined 

 to Avithin 0-3 of a division. This might allow an error of 0-12 per cent, 

 in the determination of the resistance at the centre of the bridge, and 

 15 per cent at the point farthest from the centre used in my experi- 

 ments. 



Platinizing the Electrodes. 



The electrodes, after being washed in boiling alkali and acid, Avere 

 placed in a solution prepared from a recipe given by Lummer and Kurl- 

 baum and referred to by Kohlrausch.' This solution consists of one part 



1 Wied. Ann., x. (1880), p. 326. 



2 Wied. Ann., vol. 60 (1897), p. 315. 



