and Retractions of Rods of Iron and Steel. 183 



temperature of the rod remaining constant, it has been found 

 that the rod retains the length it attained on the breaking of the 

 circuit ; that is, the rod has received with its permanent mag- 

 netic charge a permanent elongation. On passing the current 

 a second time, the rod again elongates; but the elongation is now 

 less than that which took place when the current was first passed 

 round it. On now breaking the circuit the rod retracts to the 

 length it had before the current was passed for the second time ; 

 that is, after the first magnetization and demagnetization of the 

 rod the successive elongations and retractions are equal. These 

 conditions exist until four or five subsequent make- and break- 

 circuits have been made : but now a change takes place in the 

 phenomena ; for on making the circuit the rod elongates about 

 the same as in the preceding experiments, but on breaking the 

 circuit the retraction does not equal the elongation ; so that after 

 each experiment the rod is slightly longer than before the pre- 

 ceding experiment was made. Continuing the experiments, the 

 scale gradually passes over the cross-threads ; and I have th'is 

 repeatedly caused the entire scale to traverse the field of the 

 telescope. On now allowing the rod to remain at the tempe- 

 rature it had when the current was for the first time passed 

 round it, the rod slowly retracts until, after several hours have 

 elapsed, it has the length which was observed after the first ex- 

 periment made upon it. 



Heat developed in the rod at the instant of its demagnetization. 



The above- described experiments show conclusively that the 

 minute elongations which take place on breaking the circuit are 

 due to the heat developed in the rod at the moment of its de- 

 magnetization ; for in the preceding experiments the current 

 did not heat the helix sufficiently to cause radiations from it to 

 elongate the rod ; therefore, to obtain the results described above, 

 it is important to ascertain beforehand, when the current has 

 traversed the helix for a time equal to that occupied in the ex- 

 periments, that the rod during this time does not elongate. 



If the current is sufficiently intense to heat directly the helix 

 and rod, the above phenomenon of heating on demagnetization 

 nevertheless manifests itself, and can readily be disentangled 

 from the combined effects, as will be seen further on. 



These interesting results, proving the development of heat on 

 demagnetization, were obtained a year ago without any knowledge 

 of the recent work of Jamin and Roger*; and these measures, 

 made directly on the changes in length of the rods, tend to con- 



* I have not been able to find the paper containing Jamin and Roger's 

 experiments either in the Compt.es Rendus or in the Ann. de Chim. et de 

 Phys. I have obtained the information of their results only from the fol- 



