578 DR BARTON ON TEMPERATURE VARIATION OF MAGNETITE. 



greater than this, say 2 "3 5. But to avoid over-estimating the correction to be applied 

 by this factor, let us put, for the small piece of magnetite used in the temperature 

 variation, N=2. 



25. Final Results, k and jx. — Then from curve (vi.) and the value of k' at ordinary 

 temperatures we have the values of k for all the temperatures. Hence, using the 

 relation (5), Art. 17, 



we get the values of k for the various temperatures. Finally, since ju, the magnetic per- 

 meability, = 1 + 477/c, we can calculate m as required. These results are exhibited in the 

 table on page 577, and in the curves (vii.) and (viii.). 



26. Fields used. — Note. The impressed field H' was the same throughout, namely, 

 4 - 82 c.g.s. units (nearly).* The true fields H within the specimens would be — 



(i.) For the piece of magnetite 4*9 cm. long used for the absolute value of k, 



H= r+iv^ = i+(i 4 34x-82 r 2 " 3 cg,s - units < near1 ^-* 



(ii.) For the fragment of magnetite used for the temperature variation of [x, at 

 20° C, 



H = OT = i+^216) = 3 " 4C - gS - UUitS (nearly) '* 

 (iii.) For the latter piece of magnetite at 325° C, 



4 - 82 

 H = l + (2x-342) = 2 ' 9 C ' S-S ' UnitS ( nearl y>* 



In conclusion, I have pleasure in expressing my thanks to Professor Rucker, F.R.S., 

 for suggesting this work to me ; and to Mr W. Williams, B.Sc, for winding the two 

 primary and two secondary coils which were used throughout these experiments. 



* Strictly speaking, all these fields are affected by the factor y, (see Art. 21, i.). But in all probability this factor 

 is practically unity. 



