420 REPORT — 1897. 



cylinder and the surrounding tube. In this manner it is possible to obtain 

 a very uniform temperature for the external surface, differing but little 

 from that of the circulating water. 



If the cylinder is made sufficiently thick, its temperature may he 

 approximately determined at any depth by inserting mercury thermometers. 

 It was intended at first to use thermo-couples for this purpose, but the 

 apparatus in this form would have been unsuitable for students' use in the 

 ordinary course of laboratory work, which was one of the primary objects 

 in view in the construction. It would also have been desirable to make 

 the cylinder of copper, which would have reduced the resistance of the 

 metal to the lowest point. The authors were compelled, however, to 

 content themselves for the time with cylinders of cast iron and of mild 

 steel. 



The internal surface of the cylinder, upon which the steam was 

 condensed, was a hole one inch in diameter, drilled in the solid metal. In 

 order as far as possible to minimise the resistance of the surface film of 

 condensed water, a revolving bx'ush was constructed of very thin strips of 

 steel to wipe the surface five or six times a second. This wiper was found 

 to wear in a very short time to so perfect a fit, and the water-film must 

 have been so energetically stirred, that its resistance to the passage of 

 heat must have been far less than that of the best conducting metal. 



Under these conditions, if the rate of condensation of steam were 

 infinite it should have been possible to obtain a rate of condensation many 

 times greater than the limit deduced from the cylinder condensatioii 

 experiments above mentioned. 



On making the experiment, however, it was found that the wiper 

 made very little difference to the amount of condensation. With the 

 Aviper revolving at the rate of 160 per minute, the condensation was 

 increased by about 5 per cent, on the average of several experiments. It 

 may be concluded from this that the drops of condensed water with which 

 the surface is partially covered are in such rapid motion that they do not 

 appreciably obstruct the passage of heat from the steam to the metal. 

 A film of the same average thickness, if it were absolutely quiescent, and 

 if its conductivity, as generally estimated, were only one-hundredth of 

 that of cast iron, would no doubt prove a serious obstacle, but, as a matter 

 of fact, the viscosity of water at these temperatures is so small, and the 

 motion so rapid, that the drops cannot be treated as a quiescent film. 



The temperature at various distances from the inner surface of the 

 cylinder was determined by means of mercury tliermometers inserted to a 

 depth of 8 or 9 inches in holes drilled parallel to the axis. From the 

 temperatures so observed, the conductivity of the metal and the tempei-a- 

 tures of its inner and outer surfaces could be approximately inferred. It 

 was found, however, that the presence of the holes interfered materially 

 with the flow of heat through the metal, and that the readings of the 

 thermometers under these conditions were not altogether trustworthy. 



From a number of observations on the cast-iron cylinder, a conductivity 

 of 5-3 thermal units (pound degi-ee) Fahr. per square foot per minute was 

 deduced for a gradient of one degree F. per inch ; a result which agrees 

 very closely with the authors' previous determination by a different 

 method. For the steel cylinder a conductivity of 5 '8 was similarly 

 deduced. These results apply to a mean temperature of about 140° F., 

 and are much lower than the values generally assumed for iron. 



In order to verify the authors' previous result as to the rate of 



