388 PEOF. H. L. CALLENDAR, PEEFACE TO 



The numerical values for steam were deduced in the first instance with MAXWELL'S 

 theoretical value 3 '5 for the index n in the case of a triatomic molecule, because it 

 happens, owing to the form of equation (6), that the value of n cannot be determined 

 with great accuracy from observations on the cooling effect. The value n = 3 '5 

 repi*esented most of the properties of steam satisfactorily, including the saturation 

 pressure, but did not agree very well with the observed values of the adiabatic index 

 and the specific heat, which required a value of n more nearly approaching 3'3. This 

 value was confirmed by experiments on the specific heat and the adiabatic index, 

 which were then in progress, but could not be completed in time for inclusion in the 

 original paper. The results of these experiments were first published in the article 

 "Vaporisation," in the ' Encyclopaedia Britannica/ 1902, where revised values for 

 steam were given. The methods and apparatus employed were also described and 

 exhibited at meetings of the Physical Society of London in October and November, 

 1902, but no detailed account of the work was published. 



Continuous F/oir Vacuum Calorimeter. 



The principal improvements introduced in these experiments on the specific heat 

 were as follows. The flow tube of the calorimeter, containing the heating coil and 

 the thermometers, was enclosed in a vacuum jacket to minimise external heat-loss, 

 which in turn was protected by an external steam jacket maintained at a pressure 

 slightly above atmospheric. The electric energy was measured very accurately by 

 means of a potentiometer and Weston cells, and the temperatures were observed with 

 platinum thermometers in place of couples. The chief difficulty encountered at first 

 was that of securing a perfectly steady temperature for the steam jacket, and a 

 perfectly steady current of steam, which had not been obtainable in the earlier 

 experiments in which steam had been taken from an ordinary boiler. This difficulty 

 was finally overcome by employing a boiler heated by gas, the supply of which was 

 regulated by the steam pressure itself acting on a mercury column. The regulator 

 could be set to cut off at any desired pressure within certain limits, and worked with 

 almost incredible perfection, maintaining the temperature in the jacket as nearly 

 constant as the barometric fluctuations of pressure would permit. This regulator was 

 subsequently employed by MAKOWER in experiments on the adiabatic index, and is 

 fully described and illustrated by my collaborator in the following paper. 



The method of obtaining a steady current of slightly superheated steam will be 

 readily understood from the accompanying fig. 3, which illustrates one of the many 

 arrangements adopted in measuring the specific heat. 



The steam at a steady pressure from the regulator, after passing through the 

 double-walled jacket surrounding the calorimeter, entered a separator at the top of 

 the jacket where any entrained moisture was drained off. A glass tube of small bore 

 fitted in the exit from the separator to the interior of the jacket, served as a throttle 



