' On Delicate Thermometers. 401 



for in the case of lONaCl, while the apparent volume of the salt 

 is at 20° r= 208*4, at 70° r = 219-8, falling again to r= 216'9 

 at 100°. Now, as pointed out by Ostwald*, this decrease in 

 the apparent volume of the salt does not necessarily imply a 

 contraction of the salt, but only that between 70° and 100° 

 the solution, as a whole, expands at a slower rate than that of 

 pure water ; but below 70° it expands faster, for r increases 

 from 20° to 70°. This, however, is only apparent ; for the 

 form of the interpolation-formula shows that the expansion 

 increases at a uniformly increasing rate ; the cause of the 

 apparent irregularity lies in the expansion of the water, which 

 is not uniform. 



XLVI. On Delicate Thermometers. By Spencek Umfkeville 

 Pickering, M.A., Professor of Chemistry at Bedford 

 College]. 



SOME months ago I had the honour of bringing before the 

 notice of this Society (Phil. Mag. 1886, vol. xxi. p. 330) 

 the fact that with very delicate thermometers the temperature 

 registered was never exactly the same when the column had 

 risen to the point of rest as when it had fallen to it. The in- 

 vestigation was made by placing the instrument in a large 

 calorimeter of water, removing the bulb at intervals, cooling 

 or heating it slightly, and then replacing it and observing 

 the reading. About eight separate observations were made 

 in order to determine the difference between the falling and 

 rising readings at any particular point on the stem; and 

 throughout the observations the instrument was tapped con- 

 tinuously on the upper end % to overcome the inertia of the 

 mercury in the tube. The difference w T hich was noticed was 

 explained at the time by the bulb not being of precisely the 

 same shape w r hile the mercury was being forced upwards 

 through the fine tube as when it was being dragged downwards; 

 nevertheless there were several points which rendered such an 

 explanation not altogether satisfactory. One out of the two 

 instruments examined (No. 62839) gave differences propor- 

 tional to the height of the column in the tube; in the other 

 (No. 63616) no such proportionality was noticed: moreover, 

 the instrument which showed this defect to the greatest ex- 

 tent (63616) was the one which had the smallest and strongest 

 bulb. I was subsequently indebted to Lord Rayleigh for a 

 suggestion that these differences should be attributed to the 



* Allgemeine Chemie, vol. i. p. 392. 



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



X I now employ a clockwork tapping apparatus for this purpose. 



Phil. Mag. B. 5. "Vol. 23. No. 144. May 1887. 2 E 



