500 SCIENCE PROGRESS 



The first of the two effects may be largely overcome by 

 annealing the instrument for some time at a much higher 

 temperature than that at which it is intended to be used, fol- 

 lowing this by a process of slow cooling to ordinary temperature. 

 The internal strains set up in the glass during manufacture 

 are thus relieved. Formerly it was not uncommon to find an 

 ordinary " chemical " thermometer graduated to 360° C. show 

 a permanent rise of 10° or even 20° after a brief exposure to 

 300° or 350°. Owing to the introduction of better glasses and 

 improved methods of manufacture these large permanent changes 

 are now much rarer and the temporary depressions much 

 smaller than used formerly to be the case. Another important 

 change in the modern method of constructing a mercury 

 thermometer for use above 100° C. is that, instead of exhausting 

 the space above the mercury, for work up to 350° C. this is now 

 almost always filled with nitrogen at approximately atmospheric 

 pressure. The pressure of the gas prevents the splitting up of 

 the mercury column, which often occurs above 250° C, if the 

 thermometer is vacuous. By the use of some of the newer 

 refractory glasses composed of borosilicates thermometers are 

 now constructed which will stand brief exposure to 550° C. 

 or even 575° C. To raise the boiling-point of the mercury 

 sufficiently, the filling of nitrogen or carbon dioxide gas is 

 introduced under 16 atmospheres pressure. This type of 

 thermometer, graduated usually into intervals of 2° or 5° C, is 

 generally divided from about 180° C. to 550" C. and for checking 

 purposes is provided in addition with either the freezing or 

 boiling-point. The divisions instead of being approximately 

 equal volumes throughout are made gradually shorter at the 

 higher ranges to bring the readings into accord with the gas- 

 scale. The use of these thermometers above about 480° C. is 

 however not to be recommended, as any prolonged exposure (say 

 an hour or two) even to 500° C. only, generally leads to the 

 gradual softening of the bulb, which blows out under the 

 considerable internal pressure. 



In the use of high-range thermometers the fact is often lost 

 sight of that if the whole of the mercury up to the reading be not 

 subjected to the same temperature as the bulb very considerable 

 errors may arise. In technical work very many disputes have 

 been caused by lack of attention to this point. The amount by 

 which a correctly divided thermometer would read low under 



