EXPERIMEXTAt STUDY OF HEAT LEAKAGE. 753 



an error of little consequence in total pressure drops of from 40 to 

 130 cm. of mercury when the errors due to other causes are considered. 



The orifices in the bushings which screw into the plug case, and 

 through which connection is established with the manometer, were 

 covered with several circular discs of copper gauze, capped and held 

 in place by a similar disc of sheet copper fastened to the bushing with 

 screws. The gauze discs and the plate were of a diameter about three 

 times that of the orifice, and they were located concentrically with the 

 orifice. The purpose of this arrangement was to avoid Pitot tube 

 effects. It has been used before by other experimenters. 



The possibility of moisture entering the apparatus through the 

 pressiu-e connections is obviated by the shape of the connecting tubes 

 inside the oil bath. (See Fig. 6.) 



The possibility of an error in the measurement of the difference of 

 pressure due to the fact that the pressure at the bulb of the thermom- 

 eter may not be that at the point at which the connection is taken 

 off from the plug case to the manometer, was carefully investigated. 

 It was found that, if such an error exists at all, it is of inappreciable 

 magnitude. The method employed in settling this question consisted 

 in measuring the difference in pressure between these two points 

 directly on an auxiliary manometer while the pressure difference on 

 the two sides of the plug was maintained at various values such as 

 were commonly used in regular runs, and further, in making rapid 

 and large changes in the rate of flow while noting the resulting changes 

 in the pressure difference registered by the auxiUary manometer. The 

 actual conditions of a regular run were very closely imitated by using 

 a dummy thermometer as the means of impressing the pressure 

 existing at a thermometer bulb in a regular run on one side of the 

 auxiliary manometer. The observations were made on both sides 

 of an axial flow plug such as is shown in Fig. 1. The readings of the 

 auxiliary manometer were small, averaging 0.09 cm. Hg. on the high 

 side and 0.07 on the low side. On the high side, all readings happened 

 to be zero or of the same sign, which was not that to be expected from 

 the physical circumstances of the case. On the low side, differences 

 of both signs as well as zero were observed. When the main pressure 

 difference was varied rapidly from 20.8 cm. to S6.5 cm., no change 

 was noted in the reading of the auxiliary manometer, although its 

 reading changed by 0.20 cm. when the main pressure difference was 

 brought back to 20.8 cm. There is no doubt that the actual difPer- 

 .€nces registered by the auxiliary gauge were accidental and due to 

 variations in the water level in the connections thereto. 



