550 Scientific Proceedings, Royal Dublin Society. 



An indication that the laboratory vessel has been completely 

 filled, may always be obtained by causing the mercury or water, if 

 the vessel be moist, to rise to a definite height in the capillary 

 tube g. 



When the necessary adjustments have been made, the height 

 of the mercury in the pressure tube is read off ; the difference 

 between the reading and the number corresponding to the height 

 of the division at which the measurement is made, gives the 

 pressure exerted on the gas in millimetres. The temperature of 

 the water in the cylinder C is at the same time observed. 



The analysis of the gases is conducted in a manner that is now 

 commonly adopted with this class of apparatus. To absorb the 

 carbon dioxide, a few drops of a strong solution of potash are 

 introduced into the laboratory vessel thus : — The stopcock of the 

 burette is closed, the reservoir of the laboratory vessel is fixed 

 above the level of the re-agent tube, and a small pipette contain- 

 ing the solution of potash is connected with the indiarubber tube 

 attached to the re-agent tube. As the reservoir is above the level 

 of the re-agent tube, on unscrewing the bunsen clamp, there will 

 be a tendency for the mercury to flow into the small pipette ; and 

 this may be taken advantage of in order to expel the air that 

 collects in the open end of the indiarubber tube to which the 

 pipette is attached. The air having been completely expelled, the 

 reservoir is dropped to a level slightly below that of the re-agent 

 tube, and after a few drops of potash solution have been drawn 

 into the laboratory vessel, the bunsen clamp is screwed up again. 

 The stopcock may now be opened, and the gas passed from the 

 burette into the vessel. After about ten minutes the carbon 

 dioxide may be considered to have been completely absorbed, and 

 the remaining gases may be transferred to the burette, and again 

 measured, after the mercury has been again adjusted to the levels 

 as before. The remainder of the analysis needs no further 

 description. 



When commencing to experiment with the apparatus, it was 

 anticipated that a correction for the error introduced by the bubble 

 of gas being retained in the laboratory vessel, as above described, 

 would be necessary when calculating the volume of the gas. In 

 order to determine the amount of this error, my Mend Mr. James 

 Carson very kindly made a number of experiments with the 



