652 Messrs. W. Morris Jones and J. E. Malam on the 



employed varying from 40 to 75 cm., as against 18 cm. 

 used by Owen. Auxiliary coils of german-silver wire were 

 immersed in the same ebonite oil-bath and placed in two gaps 

 of the bridge, the specimen and a standard ohm used as a 

 comparison resistance being placed in the other two gaps 

 respectively. The riders sliding on the bridge-wire were 

 fastened to continuous strings passing round pulleys placed 

 near the ends of the bridge-wire and also on a table about 

 8 ft. from the bridge, at which the person measuring the 

 changes of resistance sat. The experimenter could thus 

 operate the riders while observing the galvanometer with a 

 telescope and scale. It was found that quick and delicate 

 adjustment could be obtained by this method, the arrange- 

 ment allowing of very minute movements of the riders, to 

 which the galvanometer readily responded. The mercury 

 key used in changing riders could also be operated from 

 a distance. All resistances were placed in pure paraffin 

 insulating oil and little trouble was experienced from tempe- 

 rature variation. When a series of observations for a curve 

 was taken with steadily increasing fields it was found that 

 the balancing point on the bridge for zero field invariably 

 rose steadily through a small range (about £ s of the total 

 step for the longitudinal position) during the taking of the 

 curve, owing partly to progressive heating of the specimen,, 

 and to a small extent to hysteresis. This variation has a very 

 small effect on the percentage increase of resistance, the 

 maximum correction calculated for temperature variation 

 being about g Q of the maximum step. Any differences 

 between curves taken with fields increasing or decreasing in 

 regular sequence and those taken with high and low fields 

 alternately or following each other in any irregular manner, 

 were found to be within the limits of experimental error. 

 The platinoid compensating coil and the water-coolin <r 

 apparatus used by Owen in connexion with the electromagnet 

 were discarded. It was found that after balancing for a 

 field of 15,000 c.G.s. the current in the magnet (about 20 

 amperes) could be left running for five minutes without pro- 

 ducing any appreciable change in the zero point on the bridge, 

 the maximum galvanometer creep observed corresponding to 

 about xdoo °^ tne max i mtim st e P> whereas balancing could be 

 accomplished in much under half a minute. On the other 

 hand, if the magnet is water-cooled it is difficult to secure 

 that the temperatures in the gap between the pole-pieces 

 and at the place outside the pole-pieces where the specimen 

 is put for balancing in zero field shall be exactly the same, 

 and unsteadiness of the balancing point arises for tins 

 reason. 



