10 Messrs. C. W. Waidner and F. Mallory on 



nominal values of the coils to mean box-units at some standard 

 temperature, as determined by a previous calibration of the 

 box, were then applied, as well as the corrections to reduce 

 the bridge- wire readings to mean box-units. The resistance 

 of the platinum thermometer was then known in terms of the 

 mean box-unit. Any resistance could be measured with 

 several different combinations of coils and bridge-wire, and 

 thus the accuracy of the measurements was exposed to a 

 severe test. 



The top of the box was constructed of white marble, which 

 was found to have a very high insulation resistance. The 

 brass plugs were large and carefully ground, but at no point 

 were they larger than the top of the hole into which they 

 were inserted. This is a very important detail, as it insures 

 the impossibility of wearing a shoulder on the plug, in which 

 case no amount of pressure will give good contact. 



This box was also provided with small switches by which 

 resistances could be thrown into the battery-circuit and the 

 galvanometer shunted, which proved a great convenience in 

 finding an approximate balance. 



The bridge was always used with a key by which the 

 circuits could be closed and opened in such a way that the 

 effects of thermoelectric currents were eliminated. 



In these experiments the resistance-box was supported in 

 a large double- wall copper tank. The inner compartment of 

 this tank, in which the resistance-coils are placed, is filled 

 with a highly insulating neutral oil, which several careful 

 tests showed to be free from acid or alkali. The outer com- 

 partment was filled with water which was maintained at a 

 nearly constant temperature by a suitable thermal regulator. 

 The top of the box was enclosed in a glass case, thus pre- 

 venting sudden changes in the temperature of the top of the 

 box and the bridge-wire and the accumulation of dust. 



We take this opportunity to express our deep sense of 

 obligation to the authorities of the University of Chicago, 

 who, through the kindness of Professor W. S. iStratton of the 

 Ryerson Physical Laboratory, placed this valuable apparatus 

 at our disposal for this investigation. 



Galvanometer. — The galvanometer was of the four-coil 

 Thomson type, having what is practically a double metallic 

 case so that the disturbing effects of air-currents are reduced 

 to a minimum. The dimensions of these coils were 19 mm. 

 external diameter, 4 mm. internal diameter, 8 mm. deep, the 

 distance between centres of coils when mounted being 25 mm. 

 Each coil was wound in three sections (Nos. 36, 33, and 

 30 B. & S. wire being used) approximately according to 



