Heat across Layers of Gas. 433 



a Crookes's layer of a width of 8^ centims. would reach the 

 walls at two points, and any wider Crookes's layer would be 

 curtailed), De la Provostaye and Desains found that the rate 

 of cooling, or the escape of heat per second, which is propor- 

 tional to it, was represented by Dulong and Petit's expression 

 (which had been based on experiments made with intervals 

 of 12 and 14 centims.) until the exhaustion reached 6 millims. 

 of niercury, but that, after passing that tension, the rate of 

 cooling, instead of continuing to decrease, remained sensibly 

 constant between tensions of 6 millims. of mercury and 2-8 

 millims. (the lowest tension at which they experimented). This 

 is represented on fig. 2 

 by the horizontal line 

 from m to n. In this 

 figure the abscissas re- 

 present tensions in mil- 

 lims. of mercury, and the 

 ordinates of the thick 

 line represent the ob- 

 served rates of cooling 

 at different tensions but 

 with constant values of 

 6 1 and 02. The part of 

 the ordinate between the 

 horizontal lines repre- 



"!;ents the escanpbv radia- "^^^ shaded part is to be attributed to penetration. 



tion ; its continuation up to the parabola represents the escape 

 by convection ; and the extension upwards into the shaded por- 

 tion of the figure is due to the penetration of heat across the 

 Crookes's layer, which evidently reached the walls of the re- 

 ceiver when the tension was reduced to about 6 millims., and 

 was compressed when the exhaustion proceeded further. 



Similarly with the re- 

 ceiver of intermediate size, 

 in which a Crookes's layer 

 of a width of 4 centims. 

 would reach the walls, the 

 results obtained by De la 

 Provostaye and Desains 

 are represented graphi- 

 cally by fig. 3. In this 

 case heat leaked away by 

 penetration in appreciable 

 quantities at tensions 

 under 20 millims. of mer- 

 cury, and kept the total 



FML Mag. S. 5. Vol. 4. 









[ Fiy:'. 



) 



^^ 





n iny^ 



But 



i 

 1 



. to \ 





f 



com- 1 



ictioii\ 





Du.e 



to 



radiation. 





1 



2 



Q Tensions 





1 











No. 27. Dec. 1877. 



2F 



