246 PROCEEDINGS OP THE AMERICAN ACADEMY. 



tions were made with the spark micrometer as a measuring instru- 

 ment, and the same is true of Dr. Lodge's work with the alternate path. 

 The adaptation of the bolometer principle to this purpose furnishes a 

 much more accurate means of determining wave lengths and the 

 occurrence of resonance. It is not surprising that a change of wave 

 length of less than two per cent escaped detection by the spaik 

 micrometer method. The difference between copper and iron 

 increases as the diameter of the wires diminishes ; with wires 2 mm. 

 in diameter, the size mostly used by Hertz, the difference would be 

 exceedingly small. 



The range of wires suitable for the study of the phenomena is 

 rather limited. If the wires are larger than 1 mm. in diameter, the 

 difference between iron and copper is slight ; while with wires less 

 than 0.5 mm. in diameter the damping is so great that long wires can- 

 not be used and advantage cannot be taken of the cumulative effect 

 which is the basis of the present method. There is no disagreement 

 between the results here given and those reported by Trowbridge 

 and Bjerknes. The circuits Trowbridge used were so long that the 

 iron damped the oscillation too rapidly, and the circuits used by 

 Bjerknes were so short that the difference between the copper and 

 iron could not be determined with certainty. 



I wish to express my great obligation to Professor John Trowbridge 

 for the encouragement and suggestions that I have received from him, 

 and for his kindness in placing the resources of the Jefferson Physical 

 Laboratory so completely at my disposal. 



Conclusions. 



1. The self-induction of iron circuits is greater than that of similar 

 copper circuits under very rapid electric oscillations (115,000,000 

 reversals per second). This change in self-induction varies from 3.4 

 to 4.3 per cent in the present investigation, and increases with decreas- 

 ing diameters. 



2. The increase in self-induction produces greater damping, and a 

 shortening of the wave lengtli of between 1.5 and 2 per cent. 



3. The permeability /m of annealed iron wires under this rate of 

 alternation is about 385. 



4. For oscillations of the same period, the wave length along 

 parallel copper wires varies directly with the diameter of the wires. 

 Range of wires used 0.03915 cm. to 0.1201 cm. The maximum de- 

 crease observed is 5 per cent. 



Jeffkrsox Physical Laboratory, 

 July 24, 1894. 



