534 



SCIENCE. 



[N. S. Vol. I. No. 20. 



unfi-eqiiently happens that the dui-atiou of 

 the current is much too great, and not only 

 too great, but variable throiighout the series 

 of observations, the results of which are 

 compared. The carelessness with which 

 this method of experiment is recommended 

 by authorities who ought to know better is 

 astonishing. "We find, for instance, in one 

 of the most widely used text-books on the 

 practical application of electricity the state- 

 ment that to measure the total induction 

 across the armature of a djmamo a few 

 -turns of wu-e may be wound round the sec- 

 tion of commutation and connected in series 

 with a ballistic galvanometer, and the throw 

 of the needle, when the field circuit is closed 

 or broken, will indicate the induction. For 

 ^ny ordinary galvanometer such statements 

 are simply nonsense. 



Let us take, for the purpose of illustra- 

 tion, the measurement of the magnetic 

 quality of iron, according to Kowland's 

 method, or some one of the modifications 

 of it which have come into use. Here the 

 specimen is a ring, which, in most of the 

 recent determinations, is made up of wire 

 or thru sheet iron. The ring is surrounded 

 along its whole length by one or more mag- 

 netizing coils, and over a short length by a 

 secondary or induction coil, included in the 

 circuit of a ballistic galvanometer. The 

 inductions produced by different magnetiz- 

 ing forces are then measured by observing 

 the corresponding throws of the ballistic 

 galvanometer needle. Various modes of 

 operation are adopted, as, for instance, 

 the magnetizing force is changed by suc- 

 cessive steps from an extreme value in 

 one direction to an equal extreme in the 

 opposite direction, and then back by simi- 

 lar steps, thus passing the iron through a 

 complete cycle of magnetization. The cor- 

 responding successive throws of the galva- 

 nometer needle are then taken to indicate 

 the increased or diminished magnetic in- 

 duction, due to the different changes of 



magnetizing force. In another method the 

 magnetization is changed always fi'om the 

 extreme in one direction as the zero for 

 each observation. The change of magneti- 

 zation is in this case produced either by 

 dimtuution, and, if necessarj', reversal of 

 the magnetizing force in one magnetizing 

 coil, or by the use of a second coil and a 

 current sent through it in sucli a du-ection 

 as tends to reverse the original magnetiza- 

 tion. The reverse half of the cycle is then 

 obtained by passing the extreme current 

 through the second coil, then slowly de- 

 creasing it to the required value, and after- 

 wards suddenly breaking the circuit. The 

 changes of induction are measured as be- 

 fore by the deflections of the ballistic gal- 

 vanometer needle. Other methods might 

 be mentioned, but these will serve for our 

 present purpose. 



In order to illustrate the variable condi- 

 tions under which such experiments are 

 made, the curves given in figures 1-4 have 

 been di-awn by an autographic recorder 

 showing the actual character of the in- 

 duced current which is sent through the 

 galvanometer under different circumstances. 

 In figure 1 the numbers 1, 2, .3, 4, 5, 6, give 

 the curves of variation of current with 

 time (the ordinate being current and the 

 abscissse time) for the following set of oper- 

 ations : Two magnetizing coils being placed 

 on the iron a constant current was estab- 

 lished in one of them ; next, for curve 1, a 

 small reverse current was sent through the 

 other coil ; for curve 2, the second coU was 

 closed across the battery and the batteiy 

 cut out ; for curve 3, the battery put in cir- 

 cuit and the current again established ; for 

 curve 4, the cui-rent was increased by short 

 circuiting part of the resistance in the cir- 

 cuit ; for curve 5, the short circuit was 

 taken off and the current reduced to the 

 same valve as at the end of 3 ; for curve 6, 

 the coU was closed across the battery termi- 

 nal and the battery taken out of circuit. 



