176 



C. G. KNOTT 



site one end, is approximately zero. 



The most complete series of experiments made by Mr. Imagawa 

 was with a nickel wire of diameter 0*49 millimetres and length 56*4 

 centimetres. A current of one-third of an Ampere was kept steadily 

 flowing along it ; and the wire was subjected to a succession of 

 definite cyclic twistings, each cyclic twisting being continued until the 

 corresponding magnetic changes went through a steady succession. 

 Very complete cycles were taken for the following total to-and-fro 



twists of the wire as a whole, namely ± — , ± 7r , ± -5- , ± 2 ir , 



± Sir. These correspond respectively to twists per centimetre of 

 ± l°-6 , ± 3°2 , ± 4°-8 , ± 6°-4 , and ± 25°.6. The numbers are 

 given in Table I., and are represented graphically in Plate XXV. 



The first column on the left gives either in ordinary degrees or 

 in multiples of k the successive stages of the twisting, beginning with 

 the greatest negative twist. The second column gives, in approxi- 

 mately C. G. S. electromagnetic units, the corresponding longitudinal 

 magnetic intensities of the twisted wire from the greatest negative 

 twist to the greatest positive twist. The third column, which should 

 be read from below upwards, gives the intensities as the wire is, so to 

 speak, untwisted from the greatest positive twist to the greatest nega- 

 tive twist. The fourth column is obtained by subtracting the numbers 

 in the third column from the corresponding numbers in the second. 

 The algebraic sign of these differences shows the character of the 

 magnetic lag. Any given algebraic sign applies to all numbers below 

 it, until a number is reached which has the opposite sign. 



