PHYSICS, CHEMISTRY AND ENGINEERING 179 



K 2 (t 2 -t ) for the outside surface, since all the heat entering 

 the wall by radiation and convection per sq. ft. of inside sur- 

 face must pass through it by conduction and then be discharged 

 from the outside surface by radiation and convection. It is as- 

 sumed, of course, that the wall has come to a condition of 

 equilibrium, and is transmitting heat uniformly. Moreover, 

 the amount of heat transmitted by the wall per sq. ft. is also 

 equal to U (t-t ) where U is the unit of transmission to be de- 

 termined, already referred to, based on air temperatures. 



6. We now have four expressions for H, each of which 

 represents the heat transmission per sq. ft., and involving U, C. 

 Kj, K 2 , and the four temperatures, in which only U, t lf and t 2 

 are unknown. By elimination of t x and t 2 we find 



U= 



_+_L-hJL 



K t K 2 C 



for a simple wall, and for a compound wall in an exactly sim- 

 ilar manner we obtain 



U= j_ +J _ + x 1+ x i+ x ! _ +etc _ 



K t K 2 Cj Cj Cq 



where K x and K 2 are the respective inside and outside com- 

 bined surface coefficients, and X is the thickness in inches of 

 each material and C the corresponding conductivity. 



7. The values of K a and K 2 for a given wall material are 

 found to vary with the rate of air movement over the wall, 

 and the results of tests show that K 2 (inside or still-air coeffi- 

 cient) is practically constant so long as the air movement is 

 due to convection only. The value of K 2 increases with the 

 wind velocity, and for brick ranges from 2.38 K x at 5 miles per 

 hour to 4.22 K 2 at 20 miles per hour, with an average value 

 of 3 K 1 at 13 miles per hour, which represents our mean winter 

 wind velocity. It is therefore apparent that once this ratio is 

 determined for any building material, it is only necessary to 

 find values of C and K 2 in order to compute U for any wall. 



8. The experimental determination of the values of U can, 

 of course, be made on a limited number of wall constructions, 

 and at the same time values of C and K t can be obtained. A 

 thermal testing box as shown in Figure 2 is constructed of 

 the material to be investigated, and a heating element of high 

 resistance wire is centrally located within same. A small desk 

 fan is used to maintain a uniform temperature all over the in- 



