Twist on the Thermo-electric Qualities of Iron. 229 



apparent. In the first figure the direct and back branches 

 are distinctly separated from each other, the former lying 

 below the latter. In the second figure the distinction between 



the branches becomes less, yet their relative positions are 

 maintained. In the third figure the branches cross each 

 other, suggesting that the metamorphosis from one type to 

 the other is here or hereabout taking place. And in the 

 fourth figure we have the type originally obtained, and believed 

 to be normal. 



I cannot here help referring to a very similar feature 

 discovered in the course of Prof. Nagaoka's magnetic investi- 

 gations, called by him the " reversal of hysteresis " *. 



I may also cite one of my experiments made on another 

 piece of wire cut from the same bundle as the previous, in 

 which different ranges of twist were gone through in suc- 

 cession, and with all the intervening processes of drilling 

 omitted. It was to be expected that somewhere in the chain 

 the reversal would come into view. Suffice it to state that 

 the experiment fully confirmed the expectation. 



Experiments with still another piece, cut from the same 

 bundle and with the range of twist widened up to + 6tt and 

 even to + &tt, show that the type of the curves obtained is in 

 all cases preserved, and that the direct and back branches 

 become more and more distinct, with the two maximum points 

 separating asunder the more widely the greater are the ranges 

 of twist. 



Examination of many different samples is also here 

 wanting. 



* Journ. of the Coll. of Science, Imperial University, Tokyo, vol. iii. 

 pp. 189-207. 



Phil. Mag. S. 5. Vol. 50. No. 303. Aug. 1900. R 



