558 BELL SYSTEM TECHNICAL JOURNAL 



error is usually prohibitive and makes this type of electrode practically 

 useless. 



Sometimes electrodes of this type are amalgamated and flooded 

 with mercury and then pressed onto the sample, the excess mercury 

 being brushed away. 



This is an improvement but still leaves considerable uncertainty 

 as to the degree of contact obtained. 



Mercury Electrodes 



Primarily on account of the ease with which a liquid will conform 

 to the contour of an irregular surface, mercury has frequently been 

 used as an electrode material. The usual procedure is to float the 

 sample in a tray of mercury which serves as the lower electrode. A 

 confining ring of some form is then placed on top of the sample into 

 which a pool of mercury is poured which serves as the upper electrode. 



When transparent samples are floated in this way it is observed that 

 if a sample is simply laid flat upon the surface of the mercury, con- 



Fig. 2 



siderable air is trapped between the sample and the mercury. How- 

 ever, if the sample is lowered obliquely on to the surface of the mer- 

 cury the air can usually be eliminated. When the mercury has been 

 poured on top of the sample it is impossible to see whether the air 

 has been eliminated from the upper surface of the sample or not. 

 This is sometimes considered a serious objection to the use of this 

 form of electrode especially when used with very thin samples. It 

 may also be questioned whether or not the mercury completely fills 

 the angle between the sample and the inside surface of the ring. 



Another point to be considered in the use of this type of electrode 

 is whether the confining ring should be of conducting or of insulating 

 material. The ideal condition, of course, would be to have the con- 

 fining ring eliminated or, what is the same thing, to have it infinitely 

 thin. Practically, however, something is required to confine the 

 mercury to a definite area. Consider for a moment the arrange- 

 ment shown in Fig. 2. A is the sample under test, B is the mercury 



