544 MEASUREMENT OF RESISTANCES IN ABSOLUTE VALUE. 



of identical slits. At each oscillation of the tuning-fork, which 

 makes 127 in a second, the rotating disc may be seen; and if the 

 teeth of one circle appear at rest, it is because they replace each 



other in of a second. The observer who has the eye on the 



127 



telescope may keep the velocity absolutely constant by the simple 

 friction of the hand on one of the cords by which the motion is 

 transmitted. 



The needle, which is supported by a cocoon fibre about 130 

 centimetres in length, is protected against air currents by a glass 

 tube. This needle has an extremely small magnetic moment, so 

 that its inductive action on the frame does not come in as a term 

 of correction. In the experiments of the Committee it was a small 

 steel sphere o'Scm. in diameter and weighing about 2 grammes, 

 and its magnetisation was only about the fortieth of that which steel 

 can acquire. Its moment was equal to that which a soft iron wire, 

 10 grammes in weight, would take under the action of the earth's 

 magnetism alone. The weight of the stirrup and of the mirror 

 was relatively considerable, and the time of oscillation was about 

 10 seconds that is to say, at least 30 times that of the needle 

 alone. 



In measuring a deflection, the smallness of the needle and the 

 weak magnetic moment have theoretically no influence on the 

 exactitude of the result; but the directive action is then very 

 weak, and the torsion of the wire may have considerable influence. 

 Moreover the slightest causes, such as air currents produced by the 

 smallest changes of temperature in the bell-jar itself, which contains 

 the movable system, are sufficient to destroy equilibrium. 



It is particularly important that the magnetic axis be absolutely 

 invariable, and from this point of view the spherical shape is not 

 the best. It was chosen because a sphere uniformly magnetised 

 exerts the same external action as an infinitely small magnet placed 

 at its centre (157) ; but the same result is virtually obtained with 

 a cylinder whose length is to its diameter as \/3 is to *]z. Lord 

 Rayleigh replaces the sphere by a system of four small needles 

 0-5 cm. in 'length, which satisfy the preceding condition, and are 

 mounted on the four edges of a small cube of cork. 



1123. The experiments of the Committee present some anoma- 

 lies, the most serious of which is that the differences in the values 

 obtained amount in the mean to 3 per cent, according as the frame 

 turns in one or the other direction. This result cannot be due to a 

 previous torsion of the suspension wire ; for this torsion should have 



