EXPERIMENTAL STUDY OF HEAT LEAKAGE. 787 



regenerative action. The greater slope of the V2 (unlagged plug) 

 line is due to the inhibition of the corrective action of the regenerative 

 leakage by the flow of heat from the bath. While this line extra- 

 polates to a larger value of ju than the other, the difference is slight 

 and is of much less significance than the definite difference of slopes, 

 in view of the comparatively large deviations of the plotted points 

 from the representative straight lines, (or possibly, the genuine curva- 

 ture of both plots). 



The important conclusion from the work with these set-ups is that 

 the effect of regeneration is not great. This inference is drawn from 

 the observation that only relatively insignificant changes in heat 

 leak were produced by making changes in lagging which must have had 

 a very important effect on regeneration. This is the principal justifi- 

 cation for the belief that regenerative action is practically negligible 

 in plug S, in which heavy internal lagging is provided. 



It is to be noticed that the slopes of the upper portions of both 

 AT-Ap plots are practically identical with what has been called the 

 true jjL. 



e. Radial Flow Plug S. 



This represents the final form of the apparatus. The fx' vs. - and — 



J M 



VS. - plots both have smaller slopes than with any other plug, and 



neither exhibits any evidence of curvature. The deviations of the 

 plotted points from the representative straight lines are small, and the 



value of IX as derived from the — plot exceeds its value as derived from 



the other by only 0°.002 cm.Vkgm., or less than ^ of 1%. All 

 the points of the A T vs. Ap plot lie on a straight line, the slope of which 

 is very slightly less than the true n. The heat-leak per gram of steam, 

 which is probably due chiefly to conduction down the walls of the plug, 

 thus still increases slightly with the pressure drop, but the increase 

 is so slight that it would disappear or become a decrease if the value of 

 the 'true /x' were changed by only a small amount. There is no 

 evidence like that which, in the case of plugs VI and V2, was inter- 

 preted as pointing to regenerative heat-leak. 



The reasons for believing that the value of ju (3.182), given by the 



