114 On determining the Constant of an Electro-dynamometer. 



difference of 6 per cent, between them, the readings on mine 

 being the lower of the two. 



Thinking that this discrepancy might be due to the speed 

 at the moment of contact being greater than the average speed 

 (which was all that could be measured) , and that this difference, 

 again, might be due to the friction of the contact (as well as of 

 another contact used for counting the revolutions), I substituted 

 contact-rollers with platinum rims for the fixed wires. The 

 axles of these rollers were vertical, and were furnished at their 

 upper ends with mercury-cups into which dipped the wires 

 connected with M and N. This was to avoid thermal currents. 

 The rollers were pressed against the rim of an ebonite disk, 

 fixed coaxially to the revolving coil. Into this rim were let 

 two pieces of platinum, connected with the ends of the coil 

 and making contact with the rollers as they passed them. The 

 result of this alteration was that the difference between the 

 readings of my instrument and the tangent-galvanometer was 

 reduced to 1*5 per cent. 



I have not been able to experiment further in this matter ; 

 but I believe that, by substituting for the rollers an arrange- 

 ment of light levers, by which a bent wire could be lifted in 

 and out of two mercury-cups with each revolution of the coil, 

 thereby making and breaking contact, a still greater degree of 

 accuracy should be attainable. The addition of a fly-wheel to 

 the coil while revolving would also tend to keep the speed 

 uniform. 



A partial explanation of the disagreement of the two instru- 

 ments may be found, too, in the fact that mine is home-made, 

 and therefore not so well finished as it otherwise might be. 



With the aid of a tangent-galvanometer of known con- 

 stant, the dynamometer can be calibrated to give a conve- 

 nient absolute measure of the horizontal intensity of a magnetic 

 field. For this purpose the axis of the suspended coil must 

 be at right angles to the direction of the magnetism. 



The instrument is first set up in such a position that on 

 sending a current through the suspended coil alone, no deflec- 

 tion is produced. It is then turned bodily through a right 

 angle by means of a graduated foot, and in this position it is 

 connected in series with a tangent-galvanometer and battery. 



Let H be the strength of the field at the place where the 

 dynamometer stands, and H' the same where the tangent- 

 galvanometer stands. Then, for a current C sent through both 

 instruments and commutated, we get from the dynamometer 

 the two following equations: — 



