406 Prof. S. U. Pickering on the 



XLVII. On the Effect of Pressure on Thermometer-bulbs , and 

 on some Sources of Error in Thermometers. By Spencer 

 Umfreville Pickering, M.A., Professor of Chemistry 

 at Bedford College*. 



THE great difficulty which exists in obtaining exact con- 

 cordance between two thermometers throughout a con- 

 siderable range of their scales must have been experienced by 

 all who have had occasion to require such concordance. In 

 the course of a series of experiments, in which the tempera- 

 ture-disturbance in a calorimeter was measured simultaneously 

 with two instruments, I was much struck by the appearance 

 of a certain amount of regularity in the difference in the re- 

 sults yielded by the two instruments, which, according to 

 direct comparisons with each other through longer ranges of 

 temperature, should have been absolutely concordant. The 

 instruments being open in the scale and having large bulbs, 

 I was led to seek for an explanation of these discrepancies in 

 irregularities in the expansion of the bulbs under the pressure 

 of the column of mercury in the tube. 



The effect of pressure on a thermometer-bulb has been in- 

 vestigated by Bgen (Pogg. Ann. xi. p. 283), and by Mills 

 (Roy. Soc. Edin. xxix. p. 285), with the general result of 

 showing that the expansion experienced is directly propor- 

 tional to the pressure. But although the pressures employed 

 in Mills's experiments were considerable (ranging up to 134 

 atmospheres), the thermometers which he examined were not 

 of the most delicate character, and the coefficient of expansion 

 was but small in comparison with that possessed by most calo- 

 rimetric instruments. In the present experiments the bulb 

 of the thermometer was enclosed in a small thin brass cylin- 

 der, which was connected with a pump, by means of which a 

 vacuum, or pressure up to two atmospheres, might be pro- 

 duced. The case enclosing the bulb was previously filled 

 with melting ice and placed in a large vessel of the same, the 

 zero point, of the thermometer having been previously ad- 

 justed so as to be slightly above the level of the ice. 



Three instruments were investigated in this manner, 

 Nos. 62839, 63616, and 65108, of which the details of con- 

 struction are given in the table ii. of the preceding commu- 

 nication (see also Phil. Mag. 1886, vol. xxi. pp. 331, 340). 

 About twenty observations at different pressures were made 

 in each case, and the results are given in Table I., where are 

 entered the observed pressures and readings in rnillim.,, and also 



* Communicated by the Physical Society : read April 23, 1887. 



