Section 4/. Apparatus and Procedure. 117 



later, in attempting to make measurements with dilute sodium acetate and 

 hydrochloric acid at 218, difficulty was met with in obtaining constant 

 readings, apparently owing to adsorption by the lower electrode. The 

 platinum-black was therefore removed from it (on April 4, 1904) by rub- 

 bing it with cotton and moist pumice. Later (on June 25, 1904), in order 

 to diminish any contaminating influence of the electrode or quartz cup 

 exerted upon the small quantity of liquid within the cup, the cup and the flat 

 electrode within it were removed and replaced by a cylindrical electrode 

 of an iridium-platinum alloy with 15 per cent iridium and an insulating 

 cylinder of quartz.* This electrode was 9.7 mm. high and 7.2 mm. in 

 diameter. 



It was not found necessary to renew any of the parts of the bomb 

 throughout the course of the work ; and very little difficulty was experi- 

 enced from leaks, which occurred only a few times and were then easily 

 remedied. 



THE CONDUCTIVITY MEASURING APPARATUS. 



The conductivity was measured with an apparatus of the roller type 

 described by Kohlrausch and Holbornf and furnished by Hartmann and 

 Braun. The resistance coils of 1, 10, 100, 1,000, 10,000 ohms were of 

 manganine. Heavy copper wire leads were used to within a few cm. of the 

 slide wire and the heaters, the end connections being made of heavy flexi- 

 ble leads with brass connectors joining them to the slide wire and to the 

 leads attached to the bomb. The entire lead resistance amounted to only 

 0.02 ohm. 



The slide-wire was calibrated by the method of Strouhal and Barus, 

 once just before the conductivity work was begun and again on June 1, 

 1904, the difference in the two cases, as well as the maximum error, being 

 very slight. The corrections were, however, applied to the conductivity 

 measurements. The resistance coils were compared with standard resist- 

 ances, certified by the Deutsche phys.-technische Reichsanstalt. 



INDUCTION COILS. 



Two induction coils were used ; at first a small one of the ordinary form 

 was employed, and afterwards, in order to reduce the effect of ejection of 

 material from the electrodes, a Nernst$ string interrupter was used. The 

 quality of the minimum afforded by this latter interrupter was satisfactory 

 even with bright electrodes. In the case of all the measurements except 

 those with sodium chloride, a commutating switch was introduced between 

 the induction coil and the bridge, and the mean of the two readings taken. 



*Section 27, Part III. fLeitvermogen der Elektrolyte, 1898, p. 42, fig. 37. 



JKohlrausch and Holborn, Leitvermogen der Elektrolyte, 1898, p. 29. 



