The Comparison of Thermometers. 361 



It is necessary to protect the mercury thermometer in this 

 way, as the exposure of its bulb to the diminished pressure 

 inside the apparatus may produce a very sensible error in 

 its indications {vide Guillainne Thermoinetrie de Precision. 

 Paris, 1889). 



Regnault,* in his experiments on the Vapour Tension of 

 steam, made use of a similar arrangement. Some experi- 

 ments which he made on this point showed that the readings 

 of the thermometer at 100° were the same whether it was 

 directly exposed to the steam or not. 



The steam issuing from the opening C is condensed by 

 an ordinary Liebig's condenser. It is important that there 

 should be no narrow tubes in this part of the apparatus, 

 otherwise the pressure in the boiler may be higher than that 

 indicated on the gauge. 



The gauge (G) was of the simple barometer form, the 

 tube being 12mm. internal diameter. The height of the 

 mercury was read on a brass meter scale, placed behind the 

 tube, by a telescope at a distance of about 6ft. 



The pressure regulator D was of the form described by 

 Nicoljf and Stadel and Schummann,j and was found to give 

 satisfactory results. The pressure could be kept constant 

 within 0'5mm. for half an hour at a time. 



The pressure regulator, the gauge, and the receiver E, 

 communicate with one another through the reservoir F, 

 which consisted of a Winchester quart bottle standing in 

 a large vessel full of water. 



The first experiments were made with a glass boiling 

 apparatus, but at temperatures below 100°, the indications of 

 the thermometer were o*i° to 0*2° too high. The boiling 

 point of the platinum thermometer was found practically 



* (M^m : de I'acad : t.21.) 



t Nicol. Phil: Mag: (5) xxiii. 389. 



X Stadel and Schummann. Zeitschrift f. Instrumenten Kunde 391. 1882. 



