EXPERIMENTAT. STUDY OF HEAT LEAKAGE. 801 



are thirty-five of these in all — eight of Grindley's, thirteen of Griess- 

 mann's and fourteen of Peake's. Of the thirty-five, eighteen lie above 

 and seventeen below 3°.182 ('. cm.-/kgni. The mean of Grindley's 

 values is 8.2Sr„ at 164°.5 ('., or 3.27o at 165°, using diji/dt = - 0.031 

 cm.Vkgni.; the mean of Griessmann's values is 3.10o at 162°. 5, or 

 3.022 at Kif)"; the mean of Peake's values is 3.192 at 165°.r), or 3.208 

 at 165°. The mean of all thirty-five is 3°.179 at 164°.2 or 3.154 at 165°. 

 The actual vertical width of the band of points at 0.687 reduced 

 temperature in Davis' figure 6 is about one-third of a reduced unit, 

 or a little over 1° C. cm.Vkgm. This is about three times the differ- 

 ence between the two extreme values obtained by the present writer 

 (See Table IV). 



The precautions taken by these experimenters to avoid heat-leakage 

 do not seem in all respects adequate, in view of the writer's experiences 

 with this problem. Griessmann located his plug in a wooden pipe 

 bound with iron hoops, and untagged. The connected apparatus on 

 both sides was well-lagged. Grindley lagged his low-side steam by 

 means of a steam jacket supplied from the boiler used for the main 

 steam supply. The temperature of the jacket steam could be regu- 

 lated by means of throttle valves, but one infers from Grindley's 

 paper that it was not always practicable to make it that of the low- 

 side steam. Peake used a self-lagging arrangement on the low side 

 of his plug. His steam passed from the plug into and through a glass 

 tube, thence down on the outside of this tube. He also used thermally 

 insulating joints for his low-side pressure and flow connections, but 

 the writer's experience with similar joints is that they are of slight 

 value. Only Peake appears to have made any systematic attempt 

 to determine whether his results were in error from heat leakage. 

 His plug consisted of a mica disc with a single small hole. Copper 

 gauze was used on the low side to destroy the kinetic energy of the 

 steam jet. Peake found that throttling curves obtained with identical 

 high side conditions, but with orifices of different sizes, coincided in 

 their overlapping portions, and from this inferred that heat-leakage 

 was negligible. 'Coincidence' here must mean agreement within 

 experimental error, which is probably riot better than 0°.l or 0°.2 C. 



There is very little upon which to base a discussion of the probable 

 effects of heat-leakage on the results of Grindley and Griessmann. 

 Peake's affirmative tests for the absence of heat-leak effects can 

 scarcely be regarded as of much authority in connection with the work 

 of the other two experimenters, because of the difference between 

 his lagging arrangements and either of theirs. It would seem quite 



