532 Mr. Gr. U. Yule on the Passage of Electric 



tivities determined by ordinary methods. No absolute deter- 

 minations, however, could be made, and the measurements 

 were naturally somewhat rough. 



Since it would be important to make certain of even a 

 slight variation in the resistance with increasing rapidity of 

 oscillation, I decided last winter to take up these experiments 

 again, with such alterations as might suggest themselves to 

 render measurements more exact. After several unsuccessful 

 attempts to directly improve the procedure used by Prof. J. 

 J. Thomson, the following method was adopted : — The waves 

 were propagated along (or between) a pair of long wires 

 instead of being allowed to scatter in all directions through 

 space ; a certain length of these wires was then immersed in 

 an electrolyte, and the ratio determined in which the wave- 

 trains were weakened by their passage through this absorbent 

 layer. Measurements could be fairly accurately made with 

 an electrometer. It would have been easy from a series of 

 observations to calculate the conductivity of the electrolyte 

 used, if the change in the transmitted intensity were only a 

 consequence of absorption. But the matter proved not so 

 simple. As the thickness of the absorbent layer increased, 

 the transmitted intensity, so far from decreasing logarithmi- 

 cally, did not continuously decrease at all, but decreased and 

 increased periodically. The effect was obviously analogous to 

 the interference-phenomena of thin plates by transmitted 

 light : independently of any slight absorption a layer a 

 quarter wave-length thick transmitted a minimum, a half 

 wave-length thick a maximum, and so on. This disturbance 

 made the matter too complex to permit of an exact deter- 

 mination of the conductivity, but its intrinsic interest was 

 quite sufficient to warrant further investigation, and I give 

 in the following pages the detailed results of my experiments. 



Apparatus. 



Fig. 1 represents the actual apparatus diagrammatically. 

 The primary and secondary conductors were given the form 



Fig. 1. 



A'j L B' 



-r ! 

 cc 2 



-Z-- k e 



If, !f 4 



and dimensions previously used by Bjerknes*. The primary 

 oscillator consisted of two circular zinc plates, A A', connected 



* Wied. Ann. xliv. p. 513 (1891). 



