6 RATIO OF SPECIFIC HEATS. 



tlic ti'iiipfniturt' tliroiiiiliiiiit tlif wliolc li.itli iimst linxc Ih'cii cniistaiit I" nt 

 least, .or C. 



C. MEASUKKMKNT <>K I'ltESSt'UK,. 



In like nuumer the cleteimiiiatioii of the initial and final pressure of the 

 experimeut was pushed forward to a degree of nicety such as would not affect the 

 H values by more than 1 '/i The Hnal pressures were given by the barometer. 

 We availed ourselves of S})rung's barograph as constructed by R. Fuess. The 

 registry of this instrument is certainly correct to fractions of a millimetre, and an 

 error of as much as .2 cm. in the value of the final pressure would have influenced the 

 result ])y only 1 %. To be (juite sure as to the identity of the final pressures with 

 the atmospheric pressures for any given e.xpcrinicnt, the efflu.v tuliulure was kept 

 open to the veiy end of the measurement, /. e., until after the measurement of the 

 temperature 7a of the cooled gas. 



'i'lif initial |)ressures or pressure excesses weie registered both by a sulphuric 

 acid manometei' and by a mercury manometer of tlu' constructioii shown in figures 

 3 and 4. The columns m, (fig. 1) of the manometer were 4.5 cm. in internal diame- 

 ter. Tlie pressure tube of the mercury manometer M (fig. 1) was 1 cm. in diameter, 

 with a reservoir g about 6 cm. in diameter. This instrument was chiefly used to 

 measure pressure excesses during the course of the ex])eriment, while the mercury 

 manometer enable<l us to reduce these ])ressures to their normal value in terms of 

 tlic baroMictcr. .\ simple microscope of low power, attached to a suitable slide 

 on tilt" siilphuric-acid manometer stand, enabled us to read off the position of the 

 meniscus in terms of the attached scale. To obviate erroi-s due to parallax, an in 

 candescent nu iciirv lamp was attached to the microscope carriage and moved with 

 it. The scale being attached inunediately behind the manometer tul)e, appeared en- 

 laiged when viewed through the part of the tube filled with acid. In this way the 

 positiim of this meniscus coul<l be read off at this nianomctiT to about .02 cm., 

 conesponding to about .002;") cm. of mei-cury. F'or decreasing pressures it was 

 a<lvisable to wait after each change of measurement, until the film of acid adhering 

 to the si<les of the tuV)c had recntere<l llir Imm1\ u{ the liijiiid. 



The open end of the suli)hunca(i(l nianonictcr did not directly communicate 

 with the ail-, but opened into it through a U-tu})e, '/. (fig. I), containing calcic 

 chloride and a capillary teniiinal j>. In this way difiiculties due to absorption of 

 atmospheric moisture by the acid were avoided. 



The mercury manometer naturally required niiicli more j)recise reading than 

 the other. In.asmuch as an error of .8 ''/ in the determination of the pressure excess 

 is eijuivalenl (o an error of 1 ^ in the result, it <'learly follows that within this 



