as indicated hy their Acoustic properties. 107 



touch only by their sohd angles. If we suppose that the 

 three right lines of each system have a direction that has 

 no relation to the similar lines of adjacent systems, we shall 

 have a sufficiently correct idea of the semi-regular crystalliza- 

 tion of a mass of lead. Analogous results were obtained with 

 copper, tin, and zinc ; but the crystalline systems are more 

 extended when the metals have been kept in fusion for a long 

 time, and when they have been melted at different times. 



It is a natural consequence of this structure, that the dif- 

 ferences of elasticity of the same substance, appear in general 

 to be greater in proportion to the smallness of the diameter of 

 the circular plates employed to show them ; for the number 

 of crystalline systems which these plates contain, will be less 

 numerous as their diameters are less considerable. This is 

 actually the case. The interval between the two sounds of a 

 plate of lead, tin, or zinc, from twelve to fifteen centimetres 

 in diameter, is seldom more than a semi-tone ; while this in- 

 terval is frequently a fourth in plates of the same substances, 

 when their diameters do not exceed three or four centimetres. 

 For the same reason, a mass of metal examined by the same 

 process, will appear in general to possess a regular structure 

 in proportion to the smallness of its dimensions. 



Although it appears to be sufficiently established, that fused 

 metals are an assemblage of crystals arranged regularly, and 

 disposed in systems distant and differently inclined to each 

 other, yet it remains to be determined, how this arrangement 

 can give to these substances, properties analogous to those 

 which are observed in crystallized bodies ; but, though this 

 inquiry is a difficult one, I shall endeavour to give an idea of 

 the progress I have made in it. 



Let us take two circular plates of wood of equal thickness, 

 containing in their plane the axes of greatest and least elas- 

 ticity, and let us glue them together, so that the axes of the 

 same kind may form an angle more or less considerable ; then it 

 is obvious, that this system of crossed plates will give an idea of 

 what takes place in metals. The progress of the phenomenon 

 is then very simple, for the modes of division are very nearly 

 the same as in each of the plates taken separately, that is, that 

 one of the two is composed of two lines at right angles, and 



