314 EEPOKT— 1886. 



K, as shown in the diagram. One of the batteries contained ten Grove cells, the 

 other five, and both could be thrown in circuit in either direction by means of the 

 reversing keys M and N. By proper adjustment of the mercury cups P and Q one 

 of the batteries went in circuit, just as the dipper on the other side, coming out of 

 the mercury, threw the second battery out of circuit. By meaas of the galvano- 

 meters H and I it could be seen, during the course of an experiment, whether the 

 contacts were working properly or not. 



One of the copper wires leading to the galvanometer dipped into the bowl E, 

 the other into the trough, and could be moved up and down it in order to obtain 

 balance, the distance moved being read off a scale along the trough. There was 

 always more or less of a permanent current through the galvanometer, even though 

 the poles were both of copper ; this especially, as it was liable to sudden changes in 

 amount, caused by moving the trough pole or by accidental shaking, necessitated 

 the employment of the galvanometer in a comparatively insensitive condition. 

 Various devices were tried with a view of avoiding this current, such as electro- 

 plating the poles, using chemically pure copper sulphate, or employing for poles 

 either end of a broken wire, so as to have, if possible, metallically like poles ; these 

 were found to give an indication of current even in distilled water. 



The resistance of the smaller arm was about 800 ohms, but owing to its small- 

 ness varied greatly with the current through the change of temperature. A good 

 deal of trouble was experienced through the resistance of the smaller arm at 

 times, being less when the current passed in one direction than when it passed in 

 the other. This occurred so often that imtil its cause was understood, and steps 

 thus could be taken to prevent it, there was little hope of completing the experi- 

 ments. After various attempts it was discovered to be due to the density of the 

 solution on the outside of the beaker in the Ixiwl being slightly different from 

 that inside, so that the density in the hole, and consequently its resistance, depended 

 on the direction of the current, the electric transfusion of the liquid always changing 

 direction with the current. 



A number of experiments were made of a more or less satisfactory nature ; in 

 all no difference in the position of balance was found on reversing one of the 

 batteries. Taking the difference then to be what the galvanometer would just 

 detect, h is calculated to be less than 10"*. This is, of course, very large. The 

 limit Professor Chrystal reached in his experiments with metal conductors was less 

 than 10-i2_ 



With a view to reaching a higher limit a smaller hole was next tried. It was 

 •017 cm. in diameter, di'illed in a plate of glass '022 cm. thick. Its resistance was 

 about 2700 ohms, so that a longer tube had to be used for the large arm. The 

 balance was always found to be very different when one of the batteries was 

 changed indirection. However, this was probably entirely due to the difference in 

 temperatm-e during the small and large currents, for /( calculated from this was 

 not even as small as in the first experiments ; also the difference between the 

 balance points varied with the speed of the contact breaker. Unfortunately the 

 contact breaker, which was adjustable in its rate of vibration, had reached its 

 limit of speed, so that a completely new arrangement has now to be employed. 



On tlie Electric Resistance of Magnetite. 

 By Professor Silvanus P. Thompson, B.Sc. 



This is a preliminary note on a research begun with the view of elucidating 

 the question whether the conduction exhibited by various mineral ores and metallic 

 oxides and sulphides possessing quasi-metallic conductivity is or is not accompanied 

 by electrolysis. The substance selected, mag-netite or magnetic iron ore, is a 

 thoroughly good conductor, as is evidenced by the simple fact that if a piece of it 

 be interposed in the circuit of an ordinary electric bell the bell can be rung 

 through it. 



The sample selected was a fine homogeneous piece of ore from Arkansas, and 

 was reduced by the lapidary's wheel to parallelopipedal bar. Its total length waa - 

 5"53 centimetres, its breadth 1-52 centimetre, and its thickness 1'27 centimetre. 



