580 Dr. G. E. Allan on Magnetism of Basalt. 



proportional to the quantity of magnetite contained, should 

 lie between *01 and *05. But, since half the iron is present 

 as Fe 2 3 , the non-magnetic oxide, the lower limit should be 

 reduced to *005, the upper limit being retained, since heat 

 may change the ferric oxide into magnetic oxide. Hence, 

 on the average, and if we neglect demagnetizing action, the 

 susceptibility of basalt should lie between *005 and '05. The 

 values found by Riicker and White * lie between *0003 and 

 •0139, the average value for specimens of basalt found within 

 the British Isles being *0026. The above average limits, 

 therefore, seem to be rather high, but the average suscepti- 

 bility of the nine bars mentioned in the previous paper is 

 •0046, and that of the Stromness bar is *033. 



Effect of Hcematite. 



In its main features the magnetic behaviour of basalt 

 resembles that of magnetite ; but there are cases in which 

 another substance plays a part, as in that of the Faule Ader 

 bar (Curve 4). Folgheraiter t has described another instance 

 in which unbaked, ferruginous, non-magnetic clay was heated 

 to a dull red heat, and became noticeably magnetized by the 

 earth's field on cooling. These results seem to point to the 

 sesquioxide of iron — otherwise known as ferric oxide, hsema- 

 tite, and Fe 2 3 — as being a possible factor of disturbance in the 

 cases mentioned. MoissanJ has stated that the sesquioxide 

 of iron, when heated in an atmosphere of hydrogen or carbonic 

 acid at a temperature between 350° and 440° C, changes in 

 a few hours into the magnetic oxide Fe 3 4 , the new com- 

 pound being an allotropic form of the ordinary magnetic 

 oxide, and differing from it in density and in other particulars. 

 It is distinctly probable that this change of non-magnetic 

 sesquioxide into the magnetic oxide at about 400° C. may be 

 the cause of the peculiar changes observed, and that this 

 change constitutes an important factor, though an uncertain 

 one, in questions of terrestrial magnetism. Further, it offers 

 another means of explaining the presence of strong mag- 

 netization in surface rocks, other than that which attributes 

 it solely to the magnetizing effect of a lightning flash. 

 Pockels § employs the observed intensity of magnetization in 

 his basalt bars to calculate the magnetizing lightning current. 



* Riicker and White, Proc. R. S. lxiii. p. 460 (1898). 

 t Science Abstracts, No. 414, 1898. 

 X H. Moissan, Comptes JRendus, lxxxvi. p. 600 (1878). 

 § F. Pockels, Ann. d. PMjstk, lxiii. p. 195 (1897). 



