COMPARISON "I II. VMM M VXD (! VS TIIKIIMOM KTKKS 



53 



CALLENDAR-GUIKFITHH box was replaced by a manganine strip, cut from a large sheet. 

 Although this strip had heen carefully adjusted by filing to a very fairly uniform 

 resistance along its whole length, yet, from the method of its construction, we 

 anticipated the possibility of there IxMiig in some places mure sudden variations of 

 resistance than were likely to occur in a wire of a hard matt-rial like platinum-silver or 

 platinum-indium carefully drawn down to a certain diameter. We determined 

 therefore to substitute for the usual GAY-LussAC calibration a more complete one 

 \\itli several different "columns." As it was not always jMtssiUe to take vernier- 

 readings with the slider close up to the ends, we decided to employ only the middle 

 48 centims., and to obtain the calibration corrections for each 2 centims. over this 

 range. We are indebted to Dr. BENOIT for his advice on the best method to adopt. 

 He recommended that the whole length should Ixj divided into two parts, and that 

 tni each part a " complete" calibration should be made for every 4 centims., involving 

 the employment of " columns" of 20, 16, 12, 8, and 4 centims., and that afterwards 

 the intermediate 2 centim. points should be determined by subdivision of each 

 interval of 4 centims. into two parts. The necessary conversions of the separate 

 corrections found to one system were made exactly as in the calibration of a mercury 

 thermometer or a divided scale. 



Fig. 8. 



The method adopted for making the necessary measurements is described in a 

 paper presented to the Royal Society in 18% by one of us; the scheme of the 

 I'nimections is shown in fig. 8. 



Between the terminals P,P, of the resistance-box is connected an auxiliary adjust- 

 able resistance, having four sets of coils made of the same sample of manganine as 

 those of the bridge proper, and also a small U-shaped trough containing mercury. 

 By means of this appliance, a plan of which is shown in fig. 9, the resistance between 



