436 ME. J. H. BRINKWORTH ON THE SPECIFIC HEAT OF STEAM AT 



probably of the order 1'5. The two values of this ratio obtained with the silica 

 jacket are not reliable, since in the case of the value 2'2 when k = O'OOOSO, a change 

 in k of 100 per cent, will only cause the value of h/cm to be altered slightly, while 

 when h/cm/k = 1'2 the experiments were carried out with a partial vacuum in the 

 jacket, and there is a likelihood that the vacuum was not exactly the same through- 

 out the period of these experiments, though practically constant for all the 

 experiments made on one day. 



Discussion of the Observed Variation of the Heat-loss with the Flow. 



After completing the experiments with the non- vacuum jacketed calorimeter, the 

 actual temperature gradients in the flow-tube, both when "hot" and when "cold," 

 were observed in the case of each flow. The supply of electrical energy was adjusted 

 so that witli a distance of about 14 cm. between the end of the heating coil and the 

 thermometer the latter indicated an oiitflow temperature about 9 '3 0. above the cold 

 temperature. The curves showing the temperatures at various points along the flow- 

 tube are straight lines, provided that these temperatures are measured with the 

 middle of the thermometer, not nearer than 6 cm. nor farther away than 16 cm. 

 from the' end of the heating coil. The actual gradients fur the maximum, medium 

 and minimum flows were of the order 0'044, 0'090, and 0'170 per cm., but it is 

 necessary to point out that the third figure, though given, is not to be relied upon 

 since the total fall of temperature is so small, only amounting to 0'35 C. over a 

 distance of 8 cm. in the case of the maximum flow. The product of the temperature 

 gradient over the straight line part of the curve and the value of the corresponding 

 flow is in eacli case about 0'033. This number when multiplied by the specific heat 

 and the product divided by the mean temperature of the flow-tube, gives 0'0072, a 

 number which represents the heat-loss per cm. per degree and which is in good agree- 

 ment with that deduced from the results of the calorimetric experiments, 0'0074. 

 The gradient for a medium flow of (T365 gr./sec. is shown in fig. 10 (dotted curve.) 

 Another point worthy of notice is that the gradient ceases to be linear in that part of the 

 flow-tube near the heating coil, and the shortest distance which must exist between the 

 heating coil arid the thermometer in order to obtain complete mixing, is indicated, both 

 in the gradient observations and in the calorimetric experiments, to be about 6 cm. 

 The agreement between the values of h/cm , calculated from the calorimetric experiments 

 and from the gradient observations in the lower part of the flow-tube, indicates that the 

 &/Q term in the heat-loss represents the loss from the upper portion of the flow-tube. 

 A possible explanation of this is as follows. The steam in the flow-tube loses heat to 

 the upward current surrounding it, and heat is lost from this upward current to the 

 surrounding brass jacket. However, it is only the heat given to the jacket which is 

 actually lost. The heat lost from the steam in the flow-tube to the up-current is 

 proportional to the rise of temperature de. The temperature of the up-current is 

 consequently raised by an amount proportional to de/Q, so the actual loss to the 



