146 BRIDGMAN. 



the measurements were made, namely 8 X 10"® and 1.67 X 10~^ cm. 

 The temperature coefficient of the silver was much more nearly normal, 

 and was 0.0032, its thickness being 2.0 X 10"^. The specific resistance 

 of these metals was much higher than normal. That of the thinnest 

 gold varied from 8.4 to 19.7 X 10"^ ohms per cm. cube, average 11.6. 

 The spattered films of the same thickness varied from 15.4 to 2.3.8 X 

 10~^ average 19.2. The normal value for massive gold is 2.42 X 10~®. 

 The thicker gold had a higher resistance than the thinner, varying 

 from 9.75 to 18.5, average 13.3 X 10~^. The cause of the high specific 

 resistance of the spattered gold is doubtless to be found partly in the 

 lack of crystalline structure and perfect coherence, due to its manner of 

 formation. The high resistance of the leaf metal, on the other hand, 

 is doubtless in large part due to mechanical imperfections. Examina- 

 tion of the thinnest leaf under a microscope shows a large number of 

 folds and creases; it is practically impossible to spread the leaf on a 

 surface so that it will lie smoothly in a single unwrinkled layer. The 

 mechanical imperfections in the thicker gold were even greater than in 

 the thinner, as already mentioned, doubtless partly due to the inter- 

 ruption of the beating process at a disadvantageous stage. Two 

 samples of gold 5 X 10"^ thick had resistances of 15.0 and 10.0 X 10~^ 

 not essentially different from the other pieces. 



The specific resistance of the silver leaf varied from 3.5 to 5.1 X 

 10-^ average 4.1 X lO"*'. The normal value for silver is 1.63 X 10-^ 

 It is seen that silver lies much closer to the normal than does gold. 

 Under the microscope it too was full or minute imperfections, but of a 

 different character from the gold. There were no folds, but a number 

 of minute round perforations through the leaf. | 



It would doubtless have been most desirable if these experiments 

 could have been performed on more massive samples with the normal 

 electric constants, but the necessity of conducting away the heat seems 

 absolutely to preclude such a possibility. 



The resistance has to be artificially cooled if current densities high 

 enough to obtain an appreciable effect are to be reached, and the 

 problem of mechanical support had to be solved. For this purpose 

 the leaf metal was mounted on a piece of glass. The glass was covered 

 with a very thin coat of insulating enamel by dipping it in a very dilute 

 solution of the enamel in chloroform, the leaf metal was blown or 

 otherwise spread over the surface, and was then baked at 210° until 

 the enamel was hard. Gold or silver leaf so attached to the surface 

 of glass is full of minute flaws, but by a search under the microscope 

 parts can usually be found of sufficient homogeneity. The leaf was 



