i 



64 Mr. F. D. Brown's Notes on Thermometry. 



When the temperatures to which the thermometer is to be 

 exposed are greater than 100", the instrument should be heated 

 for some time to the highest probable temperature before the 

 steam-point is observed for the first time. In this way the 

 lowering of the zero which takes place when a thermometer is 

 heated from 100° to some higher temperature, to which it has 

 not been exposed for some time previously, is effected first of 

 all, and does not take place during the experiments, as it 

 otherwise would. 



The only objection which can be raised to this method is 

 that, when some at least of the temperatures to be measured 

 are below 100°, it is possible that the steam-point, which is 

 lowered by the first heating in steam, rises again during the 

 experiments (that is, when the thermometer is at a lower 

 temperature), and then, by the second heating in steam, is 

 again brought to the same position as at first. In this way 

 the observations in steam, although concordant, would not 



ive the true index-correction to be applied to the readings. 



hat the error which thus arises is of no importance is, I 

 think, rendered probable by the following considerations : — 

 The gradual rise of the zero of a thermometer receives its 

 most natural explanation when it is supposed that the glass 

 bulb, after having been heated and somewhat quickly cooled, 

 is in a state of strain which causes it to have a larger capacity 

 than it would have if no such strain existed. As time goes 

 on, and more especially as the thermometer is subjected to 

 small fluctuations of temperature, the particles of the glass 

 gradually yield to the forces which are acting upon them, and 

 take up new and more suitable positions. These molecular 

 movements result in a gradual diminution of the capacity of 

 the bulb, and consequently in a rise of the zero. Now it is 

 evident that, if a certain state of strain is set up when a ther- 

 mometer is cooled from 100° to 0°, when it is cooled from 100° 

 to some intermediate temperature t the strain set up will bo 

 less considerable ; there will therefore be a greater tendency 

 for the zero to rise when the thermometer is placed in melting 

 ice than when it is subjected to the temperature t. Conse- 

 quently, if it be found that, when a thermometer after being 

 heated in steam is placed in ice, no change of the zero takes 

 place for three or four hours afterwards, we may legitimately 

 conclude that, if the thermometer were maintained for the 

 same time at the temperature t , no movement of the zero would 

 occur. I have frequently kept recently-heated thermometers 

 in melting ice for several hours, renewing the ice when neces- 

 sary; and I have always observed, with all of my instruments, 

 that no change took place for the first three hours, and that 



