Draper's Barograph. 53 



the instrument would wrongly indicate a change of atmos- 

 pheric pressure. If -the cistern does not move, it supports 

 the same weight as it did before the change of temperature. 

 Hence the mercury in the fixed barometric tube extending 

 from the surface of tlie mercury in the cistern upwards, must 

 have the same weight as it had before. This weight is in part 

 balanced by the atmospheric pressure and in part rests upon 

 the fixed tube at the place near its upper end where it be- 

 comes suddenly broader. The atmospheric pressure, being 

 assumed constant, supports the same weight of mercury as 

 before; hence the fixed tube likewise supports the same 

 weight as before. But the tube supports always a definite 

 and constant percentage of the entire mass of mercury in the 

 broad part. Hence mercury cannot pass in either direction 

 between the broad and narrow parts of the tube, for such a 

 passage would alter the total amount of mercury in the broad 

 end and also the portion upheld by the tube. If these de- 

 ductions are true, then mercury in the narrow part of the 

 tube cannot expand when the temperature rises, since it is 

 impossible for the mercury to pass either into the wider part 

 of the tube above or into the cistern below. The statement 

 quoted at the beginning must therefore be wrong, for it 

 leads to an absurdity. A change of temperature in the mer- 

 cury, the pressure of the atmosphere remaining constant, 

 does affect the reading of the instrument. 



What actually takes place during a rise of temperature, 

 at constant pressure, may be briefly indicated as follows: The 

 mercury in the narrow tube begins to expand down wards as well 

 as upwards, causing a small amount to escape into the cistern. 

 The increase in the mass of the liquid in the cistern, together 

 with the lengthening of the steel springs, due to the rise in 

 temperature, will thereupon cause the cistern to fall. By the 

 several readjustments the mercury column sui3ported by the 

 atmospheric pressure is lengthened to such an extent as to 

 make the weight of the new column supported by the atmos- 

 pheric pressure equal to the weight of the old one, which 

 was shorter on account of the greater density of the mercury 

 at a lower temperature. There is now less mercury than be- 



