252 HEXRY A. EOWLAXD 



on this subject. According to the " potential theory " of electrody- 

 namics which he wished to test, unclosed circuits existed. The end of 

 one of these open circuits would exert an action on a close magnetic or 

 electric circuit. So the following experiment was made by M. Schiller, 3 

 under his direction. 



A closed steel ring was uniformly magnetized, the magnetic axis coin- 

 ciding with the mean circle of the ring. This was hung by a long fibre 

 and placed in a closed metal case. A point attached to a Holtz machin.j 

 was fixed near the box, and a brush-discharge was kept up from this 

 point. If the point acted as a current-end, a deflexion would be ex 

 pected, on the potential theory. No deflexion was observed, although 

 the calculated deflexion was 23 scale-divisions. The inference is tha', 

 either the potential theory is untrue, or else that there is no unclosed 

 circuit in this case, i. e. that the convection-currents completing the 

 circuit have an electromagnetic effect. 



Schiller's further work, not bearing directly upon convection-cur- 

 rents, leads him to the conclusion that all circuits are closed, and that 

 displacement-currents have an electromagnetic effect. 



Dr. Lecher is reported to have repeated Professor Eowland's experi- 

 ment, with negative results. His paper has not been found. 



Rontgen* has discovered a similar action; he rotates a dielectric disk 

 between the enlarged plates of a horizontal condenser and gets a de- 

 flexion of his needle. He apparently guards against the possibility of 

 this being due to a charge on his disk. A calculation of the force he 

 measures shows it to be almost one-eighth of that in the Berlin experi- 

 ment. His apparatus is not symmetrically arranged, the disk being 

 much closer to the upper condenser-plate; the distances from the upper 

 and lower plates are 0-14 and 0-25 cm. respectively. He uses a 

 difference of potential corresponding to a spark-length of 0-3 cm. 

 in air between balls of 2 cm. diameter, i. e. about 33 electrostatic 

 units, equal to the sparking potential between plane surfaces : t 0-26 

 cm. The disk is an imperfect conductor, and altogether it does not 

 seem clear, in spite of the precautions taken, that this is not diu- to 

 convection-currents. 



In the Berlin apparatus, as stated above, the needle is near the edge 

 of the disk; the magnetic effect produced is assumed to be proportional 

 to the surface-density multiplied by the linear velocity; hence the force 

 will be much greater at the edge of the disk than near the centre : but 



3 Pogg. Ann. clix, p. 456. * Sitzb. d. Berl. Akad., Jan. 19, 1888. 



