Equijpotential Lines of an Electric Current. 279 



Date. Temp. M. R. P. x 10 6 



795 ) _ 11440 



.ug, 



, S, 



1885 



21-9° ) o 



22-1° \ U 



it 



4 



a 



it 



a 



5 



a 



tt 



f 2 °o [22° 



it 



it 



6 



a 

 a 



22'° \ Zl 



a 

 a 



7 



a 

 a 



21-9° | 22 ° 



( ™ 8 lira }»««° 

 f 



792 



1716 ) 11650 



1660 f 1688 11570 



5261 KftAQ 11670 ) llftQ . 



5224 r 243 11690 f 11680 



9241 nnoc 11090 lin _ 



9131 9186 11240 1117 ° 



41) 

 31 f 



2-9° 1687 10940 



The main current through the cross was always about 0*1 abso- 

 lute unit. 



The two determinations of M on Aug. 7 were not entirely 

 independent, they being made from three sets of observations, 

 of which the second set was used with the first to give the first 

 value of M and again with the third to give the second value 

 of M. The one test at a low temperature, 2*9°, was probably 

 less reliable than the others. 



The fall of P. P. with fall of temperature appears to be 

 about J per cent per 1° C. 



The increase in the P. P. column in fields ranging from 793 

 to 5243 is perhaps purely accidental. The small but decided 

 fall in going from the field 5243 to the field 9186 undoubtedly 

 shows a true decline. The significance of this decline should not 

 be overlooked. The intensity of magnetic induction through 

 this very thin piece of metal must in all cases have been al- 

 most exactly as great as that of the magnetizing field. Hence 

 the decline in the P. P. as the intensity of the field increases 

 means a failure of the transverse effect in the cobalt to keep 

 pace with the intensity of magnetic induction through the 

 same. 



In this Journal for February, 1885, I gave as the P. P. of 

 cobalt 2460 X 10" 6 ,* with a fall of nearly 1 per cent for a fall 

 of 1° C. The specimen that gave these values was very differ- 

 ent in character from that used in the experiments of this arti- 

 cle. It was cast cobalt and was quite brittle. It probably 

 contained little, if an/, iron, but was known to contain nickel. 

 The specimen used in these later experiments is known to con- 

 tain some iron and some carbon. I do not know whether it 

 contains nickel. It is harder than most kinds of iron, but not 

 so hard as tempered steel. It is not brittle. Treatment which 

 would temper or anneal steel produced no marked effect 

 upon it. 



The impurities of the first specimen probably diminished 

 its P. P. It is not known what effect the iron and carbon 



* This is the corrected value. See explanation in this Journal for August, 1888. 



