in Nickel and Cobalt. 167 



in 



In order to compare them with the observations above given, 

 we need to plot them in some manner that will lay off the 

 values of 33 (M in Prof. Kowland's first paper) on the base- 

 line. We may then take as ordinates the values of the 

 magnetic permeability, as Rowland has done in his first paper 

 (plate ih\), or the values of k, Neumann's coefficient of 

 induced magnetization, or the values of <§, the " magnetic 

 force"* within the nickel, which would be a reversal of one 

 method used by Rowland in his first paper (plate ii.); or, 

 finally, we may use the values of 3, " the intensity of magne- 

 tization according to the German theory," as Rowland calls 

 it in his second article. 



Having plotted these various curves we may compare them 

 with (1) above, in order to determine whether our quantity 



E / 



-yr corresponds most nearly to /;-, to fc, to §, or to 3. 



The curve for /x -^ will, long before 33 has reached the 

 higher values used in the curve for ^-, have reached a maxi- 

 mum and returned nearly to the base-line. The curve for k 

 "s w iH t> e very similar to that for fi. We do not, then, 



find suggested a close connexion between fju or k, and the 

 quantity we are studying. 



The curve for «& bends upward, and is therefore quite dis- 



E' 



similar to that for ^« 



The values of 3 MB- obtained from two of Rowland's 



seriesf , made either with different specimens of nickel or with 

 one specimen under quite varied conditions, give the curves (2) 

 and (3). A separate base-line is taken for each of the three 

 curves; and the ordinates of (2) and (3) have been plotted on dif- 

 ferent scales, in order to make the general inclination of those 

 curves agree with that of (1). The values of 23, however (and 

 this is the essential particular in the plotting), are given on the 

 same scale for all three curves. The important facts about 

 the lines (2) and (3) are that they are sensibly straight for a 

 long distance, that they appear to come nearly straight from the 

 origin, and that they begin to bend perceptibly toward the 

 horizontal when 53 becomes 4000 or 5000. Although these 

 lines are carried onlv a short distance bevond this region, 



* Thomson's 'Polar Definition,' reprint, p. 307 ; and Maxwell's Treatise. 

 art. 398. 



t Phil. Mag. Aug. 1873, p. 153, and Nov. 1874, p. S'27. 



