460 Dr. J. Buchanan on the Theory of 



Or, using the values of D x , D 3 , and D 5 given immediately 

 above, 



-1-33 /V 



V ISO = I FkV. 



Again, 



Lastly, 



cx 1*214X6 V 185=1521, 

 .-. c = 1495. 



cx0-4x6" 1,33 ^T8l)~ = 468, 

 .-. c = 1586. 



cx 0-223 xe ViTo=273, 



.-. c = 1813. 



The value of c obtained from the different harmonic con- 

 stituents should obviously be the same. 



On reference to the first sentence in the discussion of this 

 curve, it will be seen that the two values of x, viz. 1*2 and 

 1*33, are comparable with one another, whilst the value of c 

 there given is only strictly comparable with the first here 

 found, viz. 1428 with 1495*. 



Curve for Hardened Iron (cf. C and No. 3, fig. 1, Part I.). 



On p. 336, Part I., 1 have given x=k, c = l, that is 

 o=1700 C.G.S. units for curve No. 3, which coincides fairly 

 well with Dr. E wing's curve. 



When the material is subjected to a cyclical variation of 

 H, harmonic analysis of Dr. Ewing's fig. 14, plate 59, he. cit. 

 for the specimen of iron which has been hardened by stretch- 

 ing, yields as the harmonic constituents 



1235 sin (£ 8 H-8D. 7 ) 



203sin(^H-31°-8); 



78sin(^|H-55°-2Y 



The amplitudes of the octave and of the third harmonic 

 are again inappreciable. 



With the same assumption regarding the value of \ as was 

 made above, I get 



Z = 7'2, ^ = 18°-3, 8 =ll°'l, 5 = 5°*1, and a?=3'4. 



