190 Mr. E. B. Rosa on the Specific 



reflexion at the mirror c through a large reading-telescope. 

 By means of a long wooden lever, attached to the torsion- 

 head a and within reach of the observer at the telescope, the 

 needle was always brought nearly to the same standard scale- 

 reading when the force was acting upon it. This insured a 

 constant distance between the needle and plate, and the 

 amount of the deflexion was determined by taking the 

 difference between this (nearly constant) deflected reading 

 and the reading when the needle and plate are brought to the 

 same potential. The distance (three metres) between the 

 mirror and scale is sufficiently great so that the force may 

 safely be taken as proportional to the scale-readings. 



The disposition of the apparatus is shown in Plate VI. 

 fig. 2. The difference of potential required is furnished by a 

 gravity-battery, any desired value less than the whole 

 potential being obtained by suitably manipulating the plugs 

 of the resistance-box. The rotating-com mutator is run by an 

 electric motor. The commutator has two ebonite wheels, 10 

 centim. in diameter, mounted on a steel axle and carrying 2 

 and 26 segments of brass, respectively, with brass contact- 

 brushes. Either wheel may be used and the speed varied 

 through wide limits, so that the rate of alternation may. have 

 any desired value from a hundred to a hundred thousand per 

 minute. Generally, however, the two-segment wheel was 

 used, and the current reversed from 2000 to 4000 times per 

 minute. A disk having 100 teeth, and gearing into the axle, 

 served as a counter, and struck a bell every 100 revolutions. 

 The potential was measured by a Weston voltmeter, inserted, 

 as shown, between the resistance-box and commutator. It 

 was carefully calibrated and, excepting an index error of 0*2 

 volt, was found correct as graduated. With respect to 

 constancy and sensibility, this instrument has performed 

 beautifully. 



As soon as the brushes of the commutator came upon the 

 insulating segments the electrodes would be discharged, 

 wholly or in part, according to the conductivity of the medium. 

 The electrodes were, therefore, short-circuited by a resistance 

 (R/, fig. 2) of about 16,000 ohms, and this insured a complete, 

 instantaneous discharge for insulating as well as for con- 

 ducting media. Thus the force acted on the needle for the 

 same relative time in all cases, the circuit being broken for 

 about ^V of the whole time when using the two-segment 

 wheel. 



III. With this apparatus, during the months of May, June, 



