Transverse Fields and Allied Phenomena. 



415 



each twist so that on making the field the transverse motion of 

 the wire was the smallest possible. Rarely did I succeed in 

 quite eliminating it. Moreover, to guard against errors of the 



preceding kind the wire was as heavily loaded as its section 

 permitted, without passing the limits of elasticity. An example 

 of the results so obtained is given in figure 6. The load sus- 

 pended from the system of thin wires was about 1700 grams. 

 The character of the observations has remained essentially the 

 same as in the foregoing experiments, and the additional care 

 bestowed has not reduced the complications. 



To test the trustworthiness of these observations and to 

 interpret them, I now rotated both wires of the system around 

 their common axis and examined them for each 90° of displace- 

 ment. The data are given in figure 7 (1-5), which is an 

 example of many similar results with other wires. To facil- 

 itate the work the twists applied lay between +90° and —90°. 

 The slopes of the lines obtained are about the same, but they 

 are alternately positive and negative. In general when the 

 deflection for no twist (fiducial zero) is positive, the slope is 

 positive and vice versa: though neither this nor the alternation 

 of sign seemed to be the invariable rule. One may note that 

 the data for the initial (0°) and final (360°) positions agree very 

 well as to slope. 



The curves of figure 7 may be interpreted in two ways : 

 either the wire is aeolotropically different in the four radial 

 directions examined, which would indicate remarkably complex 

 structure, or the wire is geometrically dissimilar, being either 

 elliptical in section or not quite straight, or both. In both 

 cases the curves obtained would be adventitious and the altera- 

 tions observed easily explained : for if the wire is virtually 

 ribbon-shaped with its plane oblique to the lines of magnetic 

 force, the tendency of the ribbon will be to set " axially " when 

 the field is made. For positions 180° apart the torques will 

 be identical ; for positions 90° apart they will vary as the sine 

 and cosine, respectively of the obliquity, and will be numeri- 

 cally opposite in sign. If the obliquity be 45° or nearly so, 

 the torques will be equal and opposite. 



