the late Dr. Joule's Thermometers. 487 



considerably from the accuracy of the comparison. After a 

 few trials which gave no satisfactory results, it was resolved 

 to carry out the comparisons in a horizontal position and to 

 determine independently the correction which has to be 

 applied in order to reduce the readings to the vertical. This 

 correction owes its origin to the expansion of. the thermometer- 

 bulb through the internal pressure of the mercury column. 

 It is determined by measuring the effect of external pressure. 

 If the addition of an external pressure p produces a rise of 

 the column of mercury equal to p{3 e degrees, and an equal 

 internal pressure produces a fall pf3 l9 then two equal pressures 

 p applied from both sides would produce a rise equal to 

 p(/3 e —/3i), but this rise may be calculated in another way. A 

 hydrostatic pressure p will alter the volume of the vessel by 

 jjKg, but the apparent contents of the vessel as measured by 

 the mercury thread will only diminish by p(fc g — K m ) where 

 Kg and K m are the coefficients of cubical compression of glass 

 and mercury respectively. 

 We thus obtain the equation 



j3i = (3 e + (fC m — Kg) 



degrees # 



=/3 e + O000154 



centim. of mercury 



This equation is deduced by Guillaume for the case of 

 thermometers with cylindrical bulbs, but, as is shown by the 

 preceding deduction, it holds quite generally. 



The apparatus used for the determination of the pressure 

 corrections is shown in PI. Y. fig. 3, and, as will be seen, is 

 almost identical with that described in Guillaume's book. The 

 thermometer is suspended in a long glass tube T, into which 

 sufficient mercury is introduced to cover the bulb ; the rest 

 of the tube is filled with glycerine in order to reduce the 

 air-space as much as possible. Side tubes, with stop-cocks 

 A and B, are led into the upper end of T — one communicates 

 with the outer air, the other with a pressure-gauge, a Win- 

 chester quart vessel K, and a water-pump. If the cocks 

 A and B are closed and the pump set to work, it will gradually 

 exhaust the vessel K, and when the pressure is sufficiently 

 reduced B is suddenly opened. Owing to the large volume 

 of K, compared to the air-space in T, the pressure-gauge 

 remains sensibly unaltered, and the reduction of pressure can 

 at once be observed on the thermometer, which is read off by 

 a kathetometer telescope. The pressure can be restored to 

 the atmospheric pressure by closing B and opening A. 



* Guillaume, Thermometries p. 103, 



