Rocks and Regional Magnetic Disturbances. 511 



or, since the currents varied inversely as the resistances of the princi- 

 pal and shunt circuits, 



where B is a constant depending on the dimensions and resistances 

 of the various parts of the apparatus. 



In one series of experiments, chosen haphazard for illustration, the 

 values of the different quantities were as follows : 



c = 17-16 cm., ^ = 40-31 cm., = 37 cm. ; 



therefore N = O0001696. 



The resistance of the principal circuit = 10'9 ohms, 



shunt = 141 



n = 30-81, n = 11'25, 



A' = 1'5836 sq. cm., A = 6'7424 sq. cm., 

 a = 13'0 cm., t\ = 17'1 cm., 



u = 22-61 cm., 



(rf 1 (Z 2 ) = 4'3 cm., (d' l d' z ) = 8'5 cm. ; 

 therefore = 0'00158. 



In each experiment a deflection was the mean of two readings taken 

 with the current direct, and reversed when the tube was in each 

 solenoid, and with the current direct and reversed in the case of the 

 auxiliary solenoid. The effect of the earth's magnetic field was thus 

 eliminated. 



Various possible sources of error had to be investigated. In the 

 first place, the suspended magnetic oxide might gradually settle in the 

 tubes, or, under the influence of the current, the particles might tend 

 to set themselves with their axes parallel to the axis of the solenoid, 

 as in Sir William Grove's well-known lecture experiment. These 

 effects were most to be feared in the case of the strong mixtures. It 

 was found that if the magnetic oxide were dried it caked and it was 

 impossible afterwards to suspend it in a state of fine and equal 

 division in the glycerine. Hence the mixtures were made by ming- 

 ling known volumes of one standard mixture of magnetic oxide and 

 water with glycerine. Thus the stronger mixtures were the more 

 aqueous, and therefore the less viscous. To investigate the possible 

 error due to settling, the following experiments were made : 



The mixtures were allowed to remain in the tube for 30 minutes : 

 no difference in the deflection could be detected in the case of the 

 weaker mixtures ; but for the strongest we obtained the following 

 results. The deflections are throughout measured in cm. 



