McKAY. — HALL EFFECT AND CURRENT DENSITY IN GOLD. 367 



This gives for ~ 0.000722. 



In these measurements the strength of the magnetic field fell off rap- 

 idly at the last. There was also a possibility that these first measure- 

 ments were affected a few per cent by short circuiting of a part of the 

 potentiometer wire. 



Another series of measurements was taken with Plate 2 of the same 

 gold. In this series the magnetizing current was read at frequent inter- 

 vals during the time of measurement of the transverse effect. This series 

 gave the following results : 



May 16. Plate 2; thickness 0.0234 cm. 



The series of measurements on the second plate is believed to be the 

 more reliable, because of the method employed to measure the magnetic 

 field, and because of the steadier conditions of the galvanometer. 



The values obtained above agree well with the value 0.00071 obtained 

 by Von Ettingshausen and Nernst * in what they considered their best 

 measurement. In the paper referred to the thickest gold used was 

 0.00392 cm. thick. The strength of the main current is not given by 

 them. 



In the remaining experiments, which were made with thinner gold, it 

 was attempted to get as great a range of primary current as possible. 

 Observations of the Hall effect were taken with weak and strong primary 

 currents alternately, except in the first set of observations given below. 



Plate No. 3 was made of gold supplied by Nickolds, of Brooklyn, 

 N. Y., and said by him to be pure ("1000 fine"). By weighing, the 

 average thickness was estimated to be about 0.00018 cm. A plate was 

 made of the same form as Plates 1 and 2, and cemented on the glass 

 plate with a mixture of resin and beeswax. The connection with the 

 electrodes was made as before by electroplating with copper. In doing 

 this, the surface of the gold was accidentally punctured at one or two 

 points, which fact, together with the uncertainty as to local thickness in 

 so thin a sheet, made the plate unfit for measurement of the absolute 

 value of the Hall effect, but did not injure it for a test of the variation of 



* Akademie der Wissenschaften, Wien, 94,2, 602 (1886). 



