FAILURE OF OHm's LAW AT HIGH CURRENT DENSITIES. 147 



cut SO as to leave a narrow isthmus of the shape shown in Figure 4. 

 It is the isthmus that carries the high current density. Connection 

 to the leaf on either side of the isthmus was by means of fine leads of 

 copper caught to the leaf with a touch of solder. A special tool had 

 to be made for cutting the isthmus. The point of a very fine needle 

 was made to travel in any desired direction across the surface of the 

 foil, scratching through to the glass, by an arrangement of two screws 



FiGUKE 4. The isthmus form of the specimen. 



at right angles to each other. With this device, under the lens of a 

 microscope, the isthmus could be cut to a high degree of precision. 

 The dimensions of the isthmus ^'aried somewhat from specimen to 

 specimen, but the length was of the order of 1 mm. and the width of 

 the order of 0.1 mm. The dimensions of each specimen were measured 

 with a microscope. 



The specimen was cooled by a stream of water flowing across the 

 isthmus at right angles to its length. This water was delivered from 

 a small glass nozzle suitably held and directed. At first kerosene was 

 used as a cooling liquid, in order to avoid danger of short circuit, but 

 the cooling was not sufficiently rapid and the desired current densities 

 could not be reached. I also tried currents of compressed air and 

 hydrogen, with results very much inferior to those even for kerosene. 

 In order to protect the specimen from the short circuiting action of 

 the water, it was covered on the upper surface, except over the isthmus 

 itself, with an additional coating of enamel. Any enamel on the 

 isthmus itself is fatal. At first I used tap water, but this was too 

 conducting. Ordinary distilled water, however, proved to be suffi- 

 ciently insulating so that no short circuiting effects from it could be 

 detected. After the distilled water had been used for some time 

 slight irregularities began to appear due to increasing conductivity 

 from miscellaneous impurities picked up from the air of the room; 

 these irregularities could be made to disappear by replacing the water 

 with fresh. 



It is necessary that the velocity of the cooling water be maintained 

 constant. For small streams, a syphon arrangement was satisfactory, 

 but for more rapid delivery the proper head was maintained by air 



