52 Notes on Barometer Construction, 



almost perfectly air free. It is again worked for a few 

 minutes as a pump ; B is now closed, and for the third time 

 the arm II. is filled up with mercury. With the naked eye 

 I could never, at this stage, even discover a small bubble of 

 air under the stopcock A, and with the aid of a lens I could 

 very seldom discern one. A is once again opened, allowing 

 a little mercury to flow through, and for greater security 

 the prescribed routine may be repeated five or six times. 

 On the last occasion of doing this the mercury is allowed to 

 escape through B only until its upper surface stands in the 

 tube at a level somewhat higher than B. Millimetre divi- 

 sions are engraved or marked on this wide portion of the 

 arm, the common zero point being at the bend. When the 

 instrument stands exactly vertical (by virtue of the adjust- 

 ing screws), then the difference of the readings of the mercury 

 columns in the two arms is identical with the real barometric 

 column. 



" Mercury can at any time be readily run off at B, or filled 

 up through C, so as to obtain a fresh upper surface of the 

 mercury in the arm in which it is exposed to the air (the 

 outer arm, II.), and regulated so as to fall within the limits of 

 the divided portion of this arm ; at the same time the per- 

 fectly air-free condition of the Torricellian chamber may be 

 proved. When this condition of perfect freedom from air 

 holds good, the uniform difference of altitude of the two 

 columns holds good, whether the mercury stands at a greater 

 or less height above B ; but should air have penetrated into 

 the vacuum chamber a slight difference of reading will be 

 found to accompany this alteration of level of the mercury 

 in II., for as the air space in the vacuum chamber is dim- 

 inished, the counteracting pressure of the air which has 

 entered it will be proportionately increased. The approxi- 

 mate compensation of capillarity is also by the same means 

 ascertained. The facile repetition of the measurement by 

 means of independent observations under the altered con- 

 ditions as above described appears to the writer to be of 

 great utility and void of all error. 



" In the first instrument constructed on this principle the 

 stop-cock A did not close quite air-tight. When the arm I. 

 was for the last time entirely filled with mercury, and when 

 the stop-cock A was closed, the author covered the latter 

 with a solution of collodion ; this provision, intended to 

 effect an air-tight joint, was found to answer admirably ; 



