284 Prof. Guthrie on the Conduction of Heat by Li 



is, that the contiguous parts of the body do not stand in identi- 

 cally the same mechanical relation to one another throughout. 

 Even two specimens of the same solid substance, alike in all 

 other respects, are probably never identical as to internal arrange- 

 ment. The solids the most homogeneous with regard to struc- 

 ture are undoubtedly those which have been derived from the 

 liquid state, either by cooling after fusion or by deposition from 

 solution. But as the physical conditions attending either of 

 these processes of solidification are never strictly uniform, we 

 cannot suppose that the result should be precisely similar in dif- 

 ferent instances. 



Still less uniformity, of course, must be expected where the 

 solid is the result of organic forces. Structure is here compli- 

 cated to the highest degree. Not only do different specimens 

 of a substance differ from one another, but the structure of one 

 and the same specimen is not alike throughout. 



Dr. J. Tyndall has shown that cubes of wood possess dif- 

 ferent conducting-powers for heat, according as the heat has to 

 travel parallel to the axis of the tree, in a radial, or in a tangen- 

 tal direction. This difference is undoubtedly due to the pre- 

 vailing cylindrical structure of the layers of which the wood is 

 formed. For if the tree's stem be supposed to be formed of 

 concentric cylinders of alternately harder and softer material, 

 and if the harder substance conduct heat better than the softer, 

 it might be foreseen that the greatest conduction will take place 

 parallel to the axis of the tree, the least radially, and that the 

 conducting-power in a tangentai direction will depend upon the 

 distance from the tree's axis at which the cube is taken. 



W\ 



Axial. 

 L 



illM 



Tangentai. 



Though less obvious, somewhat similar differences must exist 

 in different specimens of other solid substances. Thus we should 

 expect pure iron to show different conducting-powers, according 

 as it is simply solidified on cooling, or welded or rolled. For in 

 the former case it possesses a crystalline, in the second a fibrous, 

 and in the third a laminar structure. 



Gases and liquids, on the other hand, are essentially without 

 structure, and they accordingly offer the most tempting ground 

 for the examination of conduction. Two cubic inches of water 

 are physically so thoroughly identical, that deductions drawn 

 from the behaviour of the one, under given influences, most as- 



