EXPERIMENTAL STUDY OF HEAT LEAKAGE. 803 



cm.Vkgni. as compared with the \alue 3''.182 C. cm."/kgni. obtained 

 by the present writer. 



V. SUMMARY. 



1. In arriving at an experimental method of controlhng and 

 eHminating the effect of heat leakage in Joule-Thomson experiments, 

 one type of axial flow, and three types of radial flow throttling 

 apparatus, were used, and are describetl in detail. Broadly speaking, 

 an axial flow throttling apparatus may be defined as one in which the 

 direction of flow of the fluid is in general perpendicular to the leakage 

 temperature gradient, and a radial flow throttling apparatus as one in 

 which the plug, or throttling partition, is so shaped as to cause the 

 fluid to flow through it in a direction generally parallel to the leakage 

 temperature gradient. 



Attention is called to the possibility of a secondaiy leakage effect 

 (called the regeneration effect) in radial flow apparatus. The remedy 

 for this effect is internal lagging. 



2. Other apparatus, incidental to Joule-Thomson experiments and 

 used in this research, is described. Particular attention is paid to 

 the fundamental measurements of pressure and temperature differ- 

 ences, and the results of an experimental study of several important 

 details connected with these measurements are given. 



3. A short discussion of possible methods of eliminating heat- 

 leakage effects from the immediate data of a set of throttling experi- 

 ments is given. 



4. The results of a large number of adiabatic Joule-Thomson 

 experiments on superheated steam, with four axial flow plugs and five 

 radial flow plugs, are presented and discussed in detail, chiefly as an 

 experimental study of heat leakage. The best of the four particular 

 axial flow plugs used is found to be incapable of dependable results 

 without an excessive amount of experimental work. Evidence 

 pointing to the probable presence of a small regeneration effect is 

 shown to exist in the results obtained with certain radial flow plugs. 

 Affirmative evidence of the absence of error from moisture in the 

 steam is given, with a short discussion of experimental methods for 

 avoiding this difficulty. It is shown that the results obtained with 

 the final form of throttling apparatus — one of the radial flow type 

 with heavy internal lagging — may reasonably be supposed to be 

 free from appreciable heat-leak errors. Such results may also be 



