454 A. Tanakadate* on the Mean Intensity of 



would be in C.G.S. units. It thus seems that we may roughly 

 obtain the moments in C.G.S. by multiplying the numbers in 

 the y-ordinate of curves fig. 10 by 6*7, and the mean intensity 

 of magnetization by multiplying the numbers in y-ordinate 

 of curves fig. 11 by 10. 



From the curves (PL IV. figs. 1 to 5) I construct the 

 following curves of mean intensity of magnetization as a 

 function of the ratio length/diameter of the bar for differ- 

 ent magnetizing fields. From these curves we see that the 

 intensity of magnetization falls very rapidly as the ratio 

 length/diameter diminishes. As this ratio increases the 

 curves will bend toward the #-axis, and will ultimately be 

 parallel to it as long as the magnetization of ring-magnets is 

 finite. 



As a means of comparison I construct, from the curves 

 of magnetization with several wires, the diagram (PL V. 

 iig. 13) of mean intensity of magnetization as function 

 of length/v/ sectional area 4/7T, that is, what the ratio 

 length/diameter would become had these wires been moulded 

 into a single cylinder at each step instead of being separate 

 wires side by side, as in the actual case. By comparing this 

 with the curves fig. 12, we see that they both show that for 

 very flat plates the mean intensity will be extremely small ; 

 also for the case of separate wires the bend of the curve 

 occurs earlier, i. e. for less value of length/diameter. 



From the way in which the curves figs. 8 and 10 tend to 

 be asymptotic, it would seem that the induced magnetic 

 moment in a soft iron bar is not increased by increasing its 

 thickness beyond a certain proportion of its length. These 

 curves, however, are in reality obtained from experiments 

 made on a bundle of wires of given length with various 

 thicknesses of the bundle. Thus, there may be material 

 difference between this case and the case in which an actual 

 solid bar of given length is made to vary in its thickness. 

 We have no experiment of this latter case ; but if we assume 

 that, when similar bars of the same magnetic substance are 

 placed in uniform magnetic fields of the same intensity, 

 both of them will give the same amount of the mean intensity 

 of magnetization, or, in other words, if we suppose the in- 

 tensity of magnetization in a uniform magnetic field to be 

 independent of the absolute sizes of the bars, and merely 

 dependent upon the shape and nature of the magnetic sub- 

 stance, then we can trace the increase of magnetic moment 

 due to the increase of section in a bar of given length, from 



