MEASUREMENT OF DIFFERENCES OF TENSION. 183 



less (or, in other words, the negative tension is greater), 

 a fact which can be recognised by the direction of the 

 deflection exhibited by the multiplier. In order, how- 

 ever, thoroughly to understand the position of the iso- 

 electric curves, it would also be necessary to know the 

 absolute amount of the iso-electric tension at each 

 point. Instead of this, however, it is sufficient to de- 

 termine the difference between the tensions at each 

 two points, which may be found by very accurate and 

 trustworthy methods. 1 



4. To calculate these differences from the extent of 

 the deflection of the multiplier would, for reasons which 

 cannot here be further explained, be very inconvenient 

 and would afford very inaccurate results. But these 

 differences may be measured with quite sufficient pre- 

 cision by a method invented by Poggendorff and after- 

 wards improved by du Bois-Keymond. 



If it is required to determine the weight of any 

 body, the latter is placed in one of a pair of scales, 

 and weights are placed in the other until the two are 

 again in equilibrium. As in this case the action of 

 the two weights on the beam of the scales is to raise 

 each other up, they must be equal. This well-known 

 principle is, however, capable of an important generali- 

 sation. It is, for example, required to determine the 

 attraction exercised by a magnet on a piece of iron. 

 The iron is attached to one end of the beam of the 

 scales, weights to the other, till the beam is again 

 balanced. The magnet being then placed under the 

 iron, the balance of the beam is again disturbed by the 

 magnetic attraction, and weight must be added to the 

 other scale before it is restored. It is evident that the 

 1 See Notes and Additions, No. 10. 



