508 Prof. S. P. Thompson on a 



convenient in certain cases, as, for example, when the resist- 

 ance is being taken of a wire heated in an oil-bath to various 

 successive temperatures, or when some contact-resistance is 

 to be measured which must on no account be shaken. Again, 

 in Fleming's Balance the graduated wire is bent into a circle 

 and laid in a groove, contact being made by a trigger attached 

 to an arm, pivoted at the centre of the circular wire. This 

 arrangement, though I have heard no complaint of its 

 practical efficiency, is open to two objections. It must be 

 much more difficult to lay the wire with an even stress at all 

 points than is the case with a straight wire stretched between 

 two points. The range of resistances measurable by this 

 bridge is limited to something less than the actual resistance 

 of the graduated wire. Lastly, a circular scale having about 

 300 c divided into 1000 parts is not so easily obtained as a 

 straight metre-scale. 



I have therefore designed a form of resistance-balance 

 which meets the desiderata. It consists of two parts — a 

 divided wire or rheocord, and an arrangement of mercury- 

 cups, connecting-pieces, and terminals to be attached at the 

 back of the rheocord part. The figure 1, Plate XVI., shows 

 the general arrangement of these two parts ; the connexions 

 between the two parts being made by strips of copper 3 millim. 

 thick, secured under clamping-screws. X is the coil whose 

 resistance is to be measured, and S the standard coil. The 

 terminals of each of these coils dip into two large mercury- 

 cups turned in solid copper ; these are shown half size in fig. 

 2, and are marked m n and ??^ / n r respectively. Between 

 these cups lie four parallel bars of solid copper a </, b b', c c', 

 and d dl ', the ends of which terminate also in mercury-cups. 

 The bars a a! and b b 1 communicate by stout copper con- 

 nexions with the two ends of the rheocord part. The other 

 two bars communicate with the battery terminals and with 

 the terminals for the two balanced coils A and B. In all 

 accurate tests these coils should be approximately equal to 

 the X or S, and should be very nearly equal to one another. 

 As they are not required to be interchanged, they are simply 

 clamped with copper connexions under flat-headed terminal 

 screws. Short stout pieces of copper, amalgamated at the 

 ends, serve to connect the mercury-cups as shown in fig. 

 1, m being connected with a, n with d, mf with b' } and n' 

 with </. The connexions being thus made, an observation 

 is taken, and the distance x along the scale of the point of 

 contact with the galvanometer-circuit is recorded, Then the 

 connexions are altered, m being placed in communication 

 with b, n with c, w! with a', and n! with d'. This has the 



