CARBON DIOXIDE AT DIFFERENT TEMPERATURES. 215 



flow. Since the total rise in temperature in the main experiments was about 5 C., 

 it is easy to see that for the same rise in temperature in the two flows the heat loss 

 was less in the small than in the large flow by 0'4 per cent, of its total value. This 

 effect acts in the opposite direction to that ah'eady discussed, and amounts to 6 parts 

 in 10,000 on the specific heat. 



In the foregoing discussion it has been assumed that the temperature at every 

 point in a plane perpendicular to the axis of the calorimeter was the same, or, at any 

 rate, that if there was any radial temperature gradient, it was the same in both flows. 

 Now this assumption is justified up to a certain point ; the discs of copper gauze 

 referred to on p. 204 would distribute the temperature very uniformly in their 

 immediate neighbourhood ; further, the obstructions which the gas encountered in its 

 passage through the calorimeter proper were so numerous that the mixing of the gas 

 must have been very thorough, so that the heat which came from the central portions 

 of the tube to the wall must have been carried mainly by the motion of the gas, and 

 by the conduction of heat through the copper gauze, and not by pure conduction 

 through the gas. This is only another way of stating that the radial temperature 

 gradient in the gas inside the tube was zero. A considerable quantity of heat would 

 be carried through the wall of the tube by the copper gauze, but for the present we 

 will consider what would have been the state of affairs if the heat had been carried 

 entirely by the motion of the gas. This transference of heat to the wall of the tube 

 by the motion of the gas involves a sudden fall in temperature from the gas in the 

 immediate vicinity of the internal wall of the tube to the wall itself, for since from 

 each element of the surface of the tube a definite quantity of heat is being lost per 

 second, and since this heat comes from the gas, each particle of gas must fall in 

 temperature at the moment of striking the tube in order to give out the necessary 

 quantity of heat. The amount by which the gas must fall in temperature will be 

 proportional to the heat loss from the element of tube considered, and inversely 

 proportional to the quantity of gas which strikes that element per second. If the 

 gas is so perfectly mixed that each particle of it strikes the wall of the tube an 

 infinite number of times while it passes through a distance of 1 cm. parallel to the 

 axis of the tube, the fall in temperature of the wall of the tube will be zero at each 

 point ; but, obviously, the degree of turbulency of flow which is necessary merely to 

 insure a practically uniform radial distribution of temperature, or, in other words, the 

 degree of turbulency required to insure that practically all the heat which is carried 

 from the central parts of the tube to the wall is carried by the motion of the gas, will 

 fall far short of the infinite degree of turbulency cited above. 



In practice we have a degree of turbulency which is sufficiently great to insure 

 that. very little heat travels from the central portions of the tube to the wall by pure 

 conduction through the gas, and yet each particle of the gas only strikes the wall of 

 the tube a comparatively small number of times during its passage through a length 

 of tube of 1 cm. for example. It can be shown without very great difficulty that if k 



