Di;s. .1. A. HAKKKK AND I'. CIIAI'ITIS ON A 



small variations in the plug-contact resistances. On this account two new thermo- 

 meters of higher resistance were ordered from the Cambridge Instrument Co. ; then- 

 fundamental intervals were five and ten ohms respectively. They are designated in 

 this paper K. 8 and K. 9. 



Owing to the shape of the various baths in which the comparisons with the platinum 

 thermometers were made, and more especially to the difficulty of keeping dry the air 

 within the tubes of thermometers of the old form, we were obliged to modify the form 

 of the " head " of these principal thermometers. 



In the reconstruction the whole thermometer was arranged so as to be practically 

 air-tight, and the contacts were rearranged in such a manner that although the four 

 wires all left the thermometer at the same side, yet the " coil" and "compensator" 

 arms were perfectly symmetrical. At the same time the number of contacts where 

 thermoelectric effects could arise was reduced as far as possible, by suppressing the 

 brass terminals and making the connection between the platinum leads and the copper 

 fusible metal cups directly by stout copper wires, all joints being made quite secure 

 with hard solder. Though this form of thermometer-head is a little more difficult to 

 construct, we find when the four contact cups are surrounded by a thin shield of 

 polished metal to keep off air-currents, that the thermo-effects, almost invariably 

 present to some extent in the old form of thermometer, especially when rapid 

 temperature changes are progressing, are almost entirely absent. Another feature is 

 the readiness with which the glass tube can be freed from internal moisture by simply 



Fig. 7. 



connecting the small stopcock on the ebonite plate alternately to a vacuum pump and 

 to an arrangement for supplying dry air, while the whole thermometer is at a high 

 temperature. This we find to be of great importance for accurate work below 100. 

 A sketch of the thermometer in its improved form is given in fig. 7. 



XVI. STANDARDIZATION OF THE RESISTANCE-BOX. 



The standardization of the resistance-box consisted in the determination : 



( 1 ) Of the calibration corrections of the bridge- wire ; 



(2) Of the values of the resistance coils in terms of one another ; 



(3) Of the temperature-coefficient of the coils. 



XVII. Calibration of the Bridge- Wire. 

 As has previously been mentioned, the cylindrical bridge-wire employed in the 



