9i2 Dr. J. S. Gr. Thomas on Mutual Action of Convection 



smallest amount of silver solder affording a secure junction 

 to copper leads L as shown and were mounted within the 

 cylindrical brass chamber 0, the wire A passing axialiy 

 through the chamber. The ends of the chamber were 

 closed by ebonite disks, which were secured to the tube by 

 means of screws. The chamber C was contained within an 

 outer chamber formed of a brass tube screwed to ebonite 

 distance rings F. The chamber C could be rotated about a 

 horizontal axis within this stationary chamber, the latter 

 being screwed to the rigid stand G. Lugs H, affixed to 

 each of the ebonite distance rings F, served to retain the 

 ebonite ends of the chamber C in their respective vertical 

 planes during such rotation. After rotation, the inclination 

 of the plane of the wires A and B to the horizontal was read 

 by means of the pointers P shown in the elevation which 

 moved over circular scales S divided into degrees, and the 

 plane of the wares could be fixed at any desired inclination, 

 by clamping the chamber C by means of the screws P shown 

 in the sectional plan. The studs Q were inserted in the 

 ebonite ends of the chamber (), to enable the chamber to be 

 conveniently rotated. The wire A, passing axialiy through 

 the chamber, was soldered to the copper leads L which were 

 carried by the spring supports V, so that the wire was 

 always maintained just taut, the tension in the wire being- 

 regulated by means of the screws W, passing through slots 

 cut in the spring supports, the screw-heads engaging with 

 the spring as shown. The tension did not affect the resistance 

 of the wire by more than 0*1 per cent. The leads were con- 

 nected to the terminals X affixed to the respective ebonite 

 plates. The leads from the wire B passed through slots in 

 the ebonite plates and were similarly carried by the spring- 

 supports V, the tension in the wire being adjusted as already 

 explained, so that the wire remained just taut throughout the 

 experiments. The spring supports V were respectively carried 

 by racks B, R, engaging w T ith pinions P', P'. The wire B 

 could be moved to any desired distance from the wir§ A and 

 affixed thereat by tightening the screws T. The distance 

 between the wires A and B could be ascertained from the 

 reading on the scales Y, verniers carried by the racks R, R 

 enabling this distance to be determined correctly to 0*01 cm. 

 When the wires were closer together than 2*0 cm., their 

 distance apart could be very accurately determined by 

 means of a reading microscope inserted through one of two 

 openings Y in the cylindrical wall of the chamber showm 

 dotted in the sectional plan. Readings of distance apart 



