THERMAL PROPERTIES OF CARBONIC ACID AT LOW TEMPERATURES. 355 



THE USE OP THROTTLING EXPERIMENTS. 



In a throttling experiment gas flows through a throttle valve while the 

 pressures and temperatures are measured on each side of the valve. No other 

 measurements are required. It is convenient to make a number of experiments 

 starting from the same initial conditions, expanding the gas to a series of 

 lower pressures. Such a group of experiments may be called a line of experi- 

 ments, since it corresponds to expansion along one I line on the fi<j> chart. It 

 is convenient to make a number of such lines of experiments all starting 

 from the same initial pressure but from successively higher temperatures. The 

 series of final pressures to which the gas expands must be the same in all the 

 lines of experiments, if the method described below is to be used for plotting the 

 results. 



Throttling experiments cannot, without some other data, be used to plot any part 

 of the 6<j> chart, but with a single constant-pressure curve they can be used to plot all 

 the other pressure curves in the superheated area, and also the gas-limit curve. The 

 resulting curves are not affected appreciably by errors in the total heat of the liquid 

 (see Appendix III.), and are quite independent of errors in the latent heat, so that 

 throttling experiments form an independent check on the accuracy of the gas-limit 

 curve and all the constant-pressure curves. 



A method of plotting the throttling experiments so as to check the constant-pressure 

 curves is explained in Appendix IV. By this method a series of constant-pressure 

 curves is drawn which are to be compared with the pressure curves already obtained 

 from the specific heat measurements. If the curves coincide, the throttle experiments 

 confirm the accuracy of the chart ; if the curves are parallel but not coincident, the 

 throttling experiments confirm the accuracy of the specific heat measurements, but 

 indicate that there is an error in the position of the gas-limit curve, which may be 

 shifted so as to make them coincide. If they are not parallel, there is a disagreement 

 between the throttling experiments and the specific heat measurements and one or 

 other must be wrong. Since shifting the gas-limit curve will make the curves, if 

 parallel, coincide, a new gas-limit curve may be plotted derived directly from the 

 throttle experiments a method of doing this is described in Appendix V.; whether the 

 new gas-limit curve is more reliable than the original depends on the reliability of the 

 different experiments. 



In order to use the methods we have adopted the throttling experiments must 

 be arranged exactly in lines, and the series of lower pressures must be exactly the 

 same in all the lines, as already stated. When actually making the experiment 

 it is hardly possible to adjust the initial pressure and temperature and the final 

 pressure exactly to the standard values, so that the observed results have to be 

 reduced to standard conditions. A method of making this reduction is explained in 

 Appendix VI. 



3 B 2 



