114 TABLE 107.— CONDUCTION OF HEAT ACROSS AIR SPACES 



(ORDINARY TEMPERATURES) 



Loss of heat by air from surfaces takes place by radiation, conduction, and convection. 

 The two latter are generally inextricably mixed. For horizontal air spaces, upper surface 

 warm, the loss is all radiation and conduction ; with warm lower surface the loss is 

 greater than for similar vertical space. 



Vertical spaces : The following table shows that for spaces of less than 1 cm width 

 the loss is nearly proportional to the space width, when the radiation is allowed for ; for 

 greater widths the increase is less rapid, then reaches a maximum, and for yet greater 

 widths is slightly less. 



Heat conduction and thermal resistances, radiation eliminated, air space 20 cm high 



Variation with height of air space: Max. thermal resistance = 4.0 at 1.4 cm air space, 10 cm high; 

 6.0 at 1.6 cm, 20 cm high; 8.9 at 2.5 cm, 60 cm high. 



TABLE 108.— CONVECTION OF HEAT IN AIR AT ORDINARY 

 TEMPERATURES* 



In very narrow layers of air between vertical surfaces at different temperatures the 

 convection currents, in the main, flow up one side and down the other, with eddyless 

 (streamline) motion. It follows that these currents transport heat to or from the sur- 

 faces only when they turn and flow horizontally, from which fact it follows, in turn, 

 that the convective heat transfer is independent of the height of the surface. It is, accord- 

 ing to the laws of eddyless flow, proportional to the square of the temperature difference, 

 and to the cube of the distance between the surfaces. As the flow becomes more rapid 

 (e.g., for a 20° difference and a distance of 1.2 cm) turbulence enters, and the above 

 relations begin to change. For the dimensions tested, convection in horizontal layers was 

 a little over twice that in vertical. 



Heat transfer, in the usual cgs unit, i.e., calories per second per degree of thermal 

 head per cm 2 of flat surface at 22.8° mean temperature 



Where two values are given, they show the range among determinations with different 

 methods of getting the temperature of the outer plate. It will be seen that the value of the 

 convection is practically unaffected by this difference of method. 



SMITHSONIAN PHYSICAL TABLES 



