Conduct I vliy of a Solution of Copper Sidphate. 135 



necessary to ari-ange the aims so that when the resistances 

 are balanced, the niaxiimnn current shall pass through the 

 galvanometer, and generally the variable aim can be so 

 arranged that there is no perceptible deflection of the needle. 

 But in Kohlrausch's method, one cannot get complete silence 

 in the telephones, and a variation of say 1 per cent, is moie 

 noticeable in a feeble sound than in a loud one, and so 

 (unless the currents be very weak) the arms have to be 

 arranged to send the minimum current through the 

 telephones. This method also differs from the ordinary in 

 its inability to measure with accuicicy low resistances, less 

 than 10 ohms, neither can it measure very high resistances 

 more than 50,000 ohms, although, with the box I had, I conld 

 otherwise have measured 1,000,000 ohms. Where one tries 

 to measure these high or low resistances, it is found that 

 when the resistances are approximately balanced, it takes a 

 considerable alteration, say 5 per cent, in the variable arm, to 

 produce any perceptible change in the sound in the telephone, 

 and when the change is produced, it is not so much a change 

 in intensity as in quality — it almost seems like an alteration 

 in pitch, though that could not be. Besides overcoming the 

 ditHculty introduced by polarisation, there is an enormous 

 advantage in Kohlrausch's method in the way of rapidity. 

 Making an observation is the matter of seconds, instead of 

 minutes. 



The cell I used to examine the effect of change of tempera- 

 ture on conductivity was a glass tube (see Fig. 1), about 

 20 cm. in length, and 1 cm. in diameter, slightly bent. The 

 ends of tliis fitted into twi^ flat copper cups, with holes in the 

 sides, fitted with slightly conical necks. These cups were 

 about 7 cm. x 5 cm. x 2 cm. The space between the 

 glass and the copper necks was tightly packed with loose 

 hemp, and formed a perfectly water-tight joint. Wires 

 soldered to the cups gave a means of connection, the cups, 

 or rather their iuterioi- being the electrodes, the surfaces 

 exposed to the solution being about 80 square cm. The 

 cups were closed at the tops b}^ blocks of indiarubber cut to 

 fit. I had some diflticulty, however, in making tliese quite 

 water-tight, and tried several methods of stopping up the 

 cracks. It was -easy enough to stop them at ordinary 

 temperatures, but the difficulty was to find some cement 

 that did not soften at 100° C. Sealing-wax and putty were 

 among the things I tried, but neither remained water-tight 

 at 100° C. A solution of indiarubber in naptha was finally 



