SOLUTIONS AT THE FKEKZING POINT. 



exhaust pump applied should give a constant (negative) preasure. This is secured liv 

 the means shown in fig. 4. The exhaust pump is connected with one arm (a) of a 

 T-piece, the other arm (6) going to the emptying vessel, and the leg (c) passing 

 through the cork of a tall glass bottle. This cork is also 

 pierced by a second tul>e (d) open to the air above, which 

 dips a certain fixed distance below the level of the w;it-r 

 in the bottle. When the pump works, a constant stream 

 of bubbles passes through, thus keeping the pressure equal 

 to that of the water column. 



By this means liquid is sucked up from the cell into the 

 emptying vessel, and the quantity in the cell when the 

 operation is over, is found to be the same on different 

 occasions to within one or two-tenths of a gramme. Since 

 the whole quantity is more than 200 grammes, this gives 

 an accuracy greater than one in a thousand. (See p. 333.) 



Stock solution is added by means of the apparatus shown 

 in fig. 2. A quantity of solution is placed in the platinum 

 vessel and accurately weighed. According to circumstances 

 the quantity varies from "25 to 20 grammes. The long 

 platinum siphon is then inserted into the cell, and a slight 

 pressure applied by placing the india-rubber tube, already pj g 4 



described, on the neck of the vessel. The solution then 



runs over, and the last drops are ejected by another gentle blow. The filling vessel 

 is removed and reweighed, the liquid is mixed by means of the hand wheel and cord, 

 and the apparatus is ready for an observation. 



SECTION 3. On the Surface of the Electrodes. 



It is usual in experiments on the resistance of electrolytes to coat the electrodes 

 with a layer of platinum black, the result of which is to increase the effective area 

 of the plates, and so diminish polarization. In the telephone method, this platiiiiza- 

 tion is necessary, for without it silence cannot be obtained by adjusting the Wheat- 

 stone's bridge, except, perhaps, in the case of very highly resisting solutions. 



Platinized electrodes, however, are troublesome to use at high dilution. They 

 seem to have the power of extracting the salt from a solution, and of condensing it on 

 their surfaces. Thus, if a dilute solution be placed in a cell with clean platinized 

 electrodes, the resistance will rise for some time the concentration of the solution 

 seems to get less. If the solution be removed, and pure water, or a solution more 

 dilute than the first, be substituted for it, some of this occluded salt seems to come 

 out, and the resistance of the cell falls. 



Now the galvanometer used in this work is much less liable to disturbance by 



VOL. oxciv. A. 2 u 



