398 



Royal Society. 



mediately, and the position of the indices of each thermometer was 

 again read carefully ; and the water was found to be at a temperature 

 sensibly lower than before the experiment began, by about o, 6 F. 

 By this means it was proved that the forward movement of the index 

 in the protected thermometers, amounting to o, 9, was really due 

 to temperature, and not to any temporary change in the capacity of 

 the bulb produced by pressure. 



This will be rendered evident by an examination of the subjoined 

 Table of observed temperatures : — 



First Series : Pressure 1\ tons per square inch. 



Number of 

 Thermometer. 



Minimum index. 



Maximum index. 



Maximum 

 mercury. 



After. 



Before. 



After. 



Before. 



After. 



Protected ... 9645 



2 



,',' .'.'•' 5 



Mean 



Unprotected . 1 



,, o 



56 



9651 



Mean 



Temperatur 



Temperatur 



in press . 



47-0 

 47-0 



46-7 

 47-0 

 47-0 



46 ; 5 

 463 



46-4 

 46-5 

 46-0 



47-0 

 46-7 

 465 



46-5 

 46-5 

 47-0 

 46-7 



47-7 

 47-6 

 47-6 



47-6 



46-5 

 46-0 



46 

 46 

 46 



54-0 



56-5 



55-5 



118-5 



46-9 



463 



46-7 





46-1 



e of external air 



e of thermometer ] 



49 

 46-7 



49 



47 





In the Phillips's maximum thermometer, with unprotected sphe- 

 rical bulb, No. 9651, the bulb had experienced so great a degree of 

 compression as to drive the index almost to the top of the tube. In 

 all the other unprotected instruments, which had been made with 

 bulbs of unusual thickness, the index had been driven beyond its 

 proper position from 6 0, 4 to 8 0, 9 F. ; and it is obvious that the 

 amount of this error must vary in each instrument with the varying 

 thickness of the bulb and its power of resisting compression. 



Notwithstanding the great pressure to which these instruments had 

 been subjected, all of them, without exception, recovered their ori- 

 ginal scale-readings as soon as the pressure was removed. 



It will be seen that the mean rise of temperature indicated by the 

 three protected instruments was o, 9 F., whilst the mean depression 

 registered on removing the pressure amounted upon all the instru- 

 ments which admitted of its measurement to 0°'G, an agreement as 

 close as was to be expected from the conditions of the experiment. 



A second set of experiments was made upon the same set of instru- 

 ments, with the exception of 9651 ; but the pressure was now raised 

 to 3 tons upon the inch ; this was maintained for ten minutes. When 



