CONTRIBUTIONS FROM THE JEFFERSON PHYSICAL 

 LABORATORY, HARVARD UNIVERSITY. 



THE EFFECT OF LEAKAGE AT THE EDGES UPON THE 

 TEMPERATURES WITHIN A HOMOGENEOUS LAMINA 

 THROUGH WHICH HEAT IS BEING CONDUCTED. 



By B. Osgood Peirce. 



Presented December 8, 1909. Received January 5, 1910. 



In many of the determinations of thermal conductivity which have 

 been made during the last few years, the so called " wall method " has 

 been employed. That is, one face of a plate or wall of the material to 

 be experimented upon has been kept at one constant temperature for 

 a long time while the opposite face has been maintained at another 

 constant temperature, and the quantity of heat per square centimeter 

 of either face, which under these circumstances has passed per second 

 from one face to the other, has been measured in some convenient 

 way. 



In practice such a plate is of limited dimensions, and although it is 

 easy to insure that the temperatures of the faces shall be nearly uni- 

 form, it is comparatively difficult to maintain a steady gradient from 

 face to face at the edges so that the heat flow within the slab shall be 

 the same as if the faces were infinite in extent. If, however, the faces 

 of the specimen to be used are small enough, it is possible to prevent 

 almost entirely the escape of heat at the edges by surrounding the 

 periphery by an arrangement like a Dewar flask. This is impracticable 

 when for any reason the plate has to be large, and in this case it is 

 necessary to make the thickness of the wall so small compared with the 

 dimensions of the faces that the lines of flow of heat from face to face 

 in the central portion of the slab shall not be appreciably distorted by 

 loss of heat through the edges of the wall. 



Some time ago, in an attempt to obtain an accurate average value 

 of the conductivity of a given stratum in a certain deep mine, I had 

 occasion to apply the wall method to some blocks of stone which were 

 not perfectly homogeneous, and in order to represent the material fairly 

 it seemed best to use a slab eight centimeters thick for each determina- 

 tion. The slabs were square and the edges were covered with lagging 



