14 Dr. F. Auerbach on the Passage of 



slight dependence of the resistance on the magnetization of 

 the English steel, or whether considerably stronger magneti- 

 zing forces were not required here in order to produce, even 

 approximately only, the state of so-called saturation, I made 

 use of a Gramme machine of the newer construction (1875), 

 a short time previously acquired for this laboratory by Prof. 

 0. E. Meyer, and, driven by a steam-engine, capable of ope- 

 rating with considerable effect. Its distance from the galva- 

 nometer amounted to about 25 metres ; the electromagnet 

 was set up between the two, at the distance of 5 metres from 

 the machine, and 20 from the galvanometer, and connected in 

 both directions by thick copper wires. The spiral never grew 

 so hot that it could not be touched with impunity. The caout- 

 chouc pipe never exhibited any trace of heating, even with the 

 most rapid rotation of the machine (1200 revolutions per 

 minute). Notwithstanding, many of the series of experi- 

 ments made with soft iron wires, especially with thin ones, 

 proceeded in a manner which I can only account for by influ- 

 ences of temperature. If we seek to eliminate these, we ob- 

 tain increases of resistance by magnetization rising to 1 per 

 cent, and over. Simpler and more reliable are the results of 

 the experiments with steel wires, the mean values of which 

 are here extracted : — 



27. H = 1D; M = Gramme machine (number of revolu- 

 tions,^ =700); w=110. Thin steel wire F 5 , 1 = 2000, d = 0'21. 



w =8-4960 w=8-4803 S= -00019. 



28. The same repeated (4 series of experiments). 



Wo = 8-6415 w= 8-6346 S= -00008. 



29. Steel wire F 3 . H = 1D; M= Gr. machine (p = 800). 

 ^ o =0-3435 20=0-3419 w = 0-3435 £=-00047. 



30. To this experiment was joined immediately a control 

 experiment with a copper wire. (Such control experiments 

 were also previously made from time to time.) H = 1D; 

 M= Gr. machine (p = 1000-1200); Z=2220; d = 0'70. Mean 

 values : — 



w =0-14138 2^ = 0-14149 w = 0'14144 S= +000056. 



For copper wire, therefore, $ is, first, positive, and, secondly, 

 much smaller in absolute value than in 29, although p is much 

 greater. Besides, all the other experiments with copper wires 

 have yielded considerably lower values still of h. If these 

 alterations of resistance should prove to be connected with the 

 magnetic behaviour of copper, they will disprove the generality 



