1886.] 



Magnetisation in the Length of Iron Wires. 



261 



disturbance thus introduced would of course be altogether insensible ; 

 but in making measurements in which a hundred-thousandth of a 

 millimetre is a considerable quantity, it is far from negligible, as indeed 

 was sufficiently proved by the inconsistency of the results obtained 

 in some of my earlier experiments when the wire was free and the 

 coil attached to the table T.* After the coil had been suspended 

 upon the wire all such inconsistency at once disappeared, for no inter- 

 action between the two could then produce any external effect. 



Since the apparatus was not calculated to bear any very heavy 

 weight it was necessary to use wires of small sectional area. Thin 

 wires moreover possess an advantage in becoming more strongly 

 magnetised by a given current than thick wires of the same length. 



The results of a series of experiments are presented in a synoptical 

 form in the subjoined Table. Four specimens of iron were used. 

 The first was a wire of commercial iron, 1*2 mm. in diameter, which 

 had been softened by heating in a gas flame ; the second was a strip 

 of annealed charcoal iron, 5'5 mm. wide and 0'55 mm. thick, its 

 sectional area being about 3 mm.; the third was a piece of hard 

 unannealed wire, 2*6 mm. in diameter ; and the last was a wire of 

 very pure soft iron, 3' 25 mm. in diameter, which had been carefully 

 annealed. These were successively fixed in the apparatus, and 

 loaded with weights varying from 3 lbs. — that of the coil and lever 

 alone — to a total of 14 lbs. While under the influence of each load, 

 four observations were made in the case of each wire : (1) A deter- 

 mination was attempted of the smallest magnetising current which 

 sensibly affected the length of the wire in the direction of elongation 

 or retraction. (2) The current producing maximum elongation (if 

 any), and the extent of such maximum elongation were found. 

 (3) A determination was made of the critical current which was 

 without effect upon the original length of the wire, i.e., the current 

 of such strength that a weaker one would cause elongation and a 

 stronger one retraction. (4) The retraction produced by a fixed 

 current of 1*6 ampere was measured. 



The first operation, that of finding the smallest current which pro- 

 duced a sensible deflection, was not easy to perform satisfactorily. 

 Small differences in the disposition of the lever and mirror might 



* The same source of error troubled me much in the experiments described in 

 my former paper until I adopted a similar method of avoiding it. The apparatus 

 used by Joule was far larger, more massive, and presumably less delicate than mine. 

 In the instrument employed in his stretching experiments the lever alone without 

 any additional weight produced a tension in the wire of 80 lbs. Errors arising 

 from solenoidal suction would therefore be less sensible, but it is difficult to believe 

 that pome of his results were not affected by them, especially in the case of the 

 experiment (No. 8) on hard steel described at p. 245 of the Eeprint, which I believe 

 no one has succeeded in repeating. 



VOL. XL. J 



