222 MR. W. F. G. SWANN ON THE SPECIFIC HEATS OF AIR AND 



and it is satisfactory to observe that this order of accuracy is approached even in 

 the case of Experiment 8, and the mean of Experiments 15 and 16, and these 

 experiments took no part in the determination of the constants of the equation. The 

 full account of the method by which the equation was deduced is contained in the 

 archives. 



The percentage decrease in Q per degree increase in t varies with H ; it can be 

 shown from the above table, combined with the empirical equation, to amount to 

 G'385 per cent, for H = 29 cm. and 0'444 per cent, for H = 10'59 cm. To obtain the 

 percentage decrease for intermediate values of H, it is sufficiently accurate to assume 

 a linear variation with the flow. It is interesting to observe that the linear relation 

 gives 0'5 for the percentage decrease when the flow is zero, and this is not far from 

 the theoretical value calculated by neglecting the inertia effects, which value we 

 should of course expect to obtain for indefinitely small values of the flow. 



(18) The Main Calorimetric Experiments. In the experiments at air temperature 

 water was siphoned from the tank N, fig. 2, round the double walls of the calorimeter 

 jacket into another vessel, from which it was poured back again into the tank, where 

 it was kept well stirred. The apparatus was tested for leaks, and the pump was 

 started. The temperature of the gas was noted by means of the thermometer K 

 when it had become steady, i.e., after about half-an-hour. The thermometers were 

 then arranged differentially, and the balance point on the bridge was again deter- 

 mined. The electric current was then started through the heating coil, and the 

 rheostat was adjusted so that the rise in temperature was about 5 C. or 6 C C. ; the 

 current remained almost perfectly steady when the various resistances in the circuit 

 had warmed up to their final values. The screw stirrer in the tank containing the 

 tubes was started, and in about 20 minutes the balance point on the bridge became 

 steady to about 0'2 mm., i.e., to about 0'005 C. ; but the final readings were not 

 taken until about half-an-hour after the commencement of the heating, by which 

 time the balance had usually become steady to O'l mm. or less. The clip A, fig. 2, 

 was adjusted so that the bell of the pressure regulator was quite still, or moving so 

 slowly that it took about 10 minutes to rise or fall two or three centimetres in order 

 to avoid fluctuations in pressure. This adjustment could be performed with com- 

 parative ease after a little practice. The bridge reading, potentiometer readings 

 across the secondary standard and across the heating coil, and the gauge reading 

 were then observed over a period of three or four minutes in order to make sure that 

 no fluctuations were taking place, or that, if there were any very slight fluctuations, 

 they were consistent with one another. The temperatures of the gauge, tubes, 

 manganin coil, and cadmium cells were then noted, and the potential and current 

 leads were quickly interchanged (see p. 204). While the temperature conditions 

 were becoming constant, the potentiometer reading across the cadmium cells was 

 taken, also the reading of the barometer and its temperature. When the tempera- 

 ture rise had again become steady, the bridge gauge and potentiometer readings 



