434 



Mr. G. J. Stoney on the PenetrMion of 



escape of heat nearly constant between tensions of 20 and 4 

 millims. 



12. But the most decisive experiments were made with the 

 cylindrical vessel, which was the smallest of the three receivers. 

 In it the interval between the bulb and the walls of the 

 vessel was only 2 centims. 



With this vessel the rate of cooling was slower than in the 

 two larger receivers at all tensions from 760 millims., or the 

 tension of an atmosphere, down to about 45 millims. This 

 seems to indicate that the convection currents were impeded 

 by the form and small size of the cylinder; so that if the or- 

 dinates of the parabola r represent the rate of cooling which 

 would result from convection in a large vessel, the ordinates 

 of some lower curve such as Ow will represent the rate due 

 to convection in the cylinder. The observations recorded by 

 De la Provostaye and Desains enable us to fix the points ?>^, r, 

 and 7i, corresponding to the tensions 70, 45, and 15 millims., 

 nearly in a hoi'izontal line. They also state that near the ten- 

 sion of 6 millims. the rate of cooling diminished with '' exces- 

 sive rapidity," but that nevertheless at a tension of 2*8 millims. 

 it still exceeded by a large amount that which presented itself 

 at the same tension in their largest receiver, and which is re- 

 presented in fig. 2. These statements indicate that the obser- 

 vations, if plotted down, would have given a curve like the 

 thick line of fig. 4. It is hardly necessary to point out that 





Due^ to 



convection. 



the larger development of the phenomenon, and its exhibiting 

 itself at higher tensions with each diminution of the size of 

 the receiver, as shown in the foregoing diagrams, are in the 



