production of the Prismatic Structure of Basalt. 129 



or to the numbers 100, 87*7, 81'6. It thus appears that, 

 amongst the three figures, the effort producing rupture is the 

 greatest in the hexagon, while the resistance to rupture is the 

 smallest ; it is thus evident that the work done to produce frac- 

 ture in the last-named figure will be much less than in either 

 of the other two. Calling r the resistance to splitting in the 

 respective cases and R the splitting or rending efforts, it is ob- 

 vious that p affords a measure of the relative work required for 

 splitting, or 



unity for the triangle, 



:0*68 for the square, 

 : 0*5 19 for the hexagon, 



r 



100 



R 



~~ioo : 





87-7 



>> 



129 





81-6 



157 



which is thus proved to require little more than one half the 

 effort that would be necessary for division into equilateral 

 triangles. 



Splitting-up once commenced in this form at the cooling sur- 

 face, must, as the cooled couche thickens, proceed deeper and 

 deeper into the mass in a direction perpendicular to the surface 

 of cooling, dividing the whole into hexagonal prisms, the lower 

 terminations of the planes of fracture between which can at any 

 moment never reach deeper into the mass than the distance at 

 which it has cooled downwards from its original temperature to 

 the " splitting-temperature," or to that at which the stiff viscous 

 mass still existing deeper down shall have assumed sufficient 

 rigidity to break when exposed to tension instead of yielding to 

 it as a plastic mass. We have now to fix, as far as existing in- 

 formation will enable us, what is this superior limit of tempera- 

 ture at and below which splitting by contraction can commence 

 and continue. And here, as in every other portion of the sub- 

 ject before us in which numerical data would be of the very 

 highest importance, we have to lament their almost total absence. 

 None of those who have speculated, more or less vaguely and in 

 many words, upon the subject have ever thought it worth 

 while to determine experimentally the dilatation by heat 

 (i. e. the coefficient of contraction) of any one sort of basalt, or 

 its extensibility at any temperature when exposed to tensile 

 strain, or its total extension at rupture by such strains. Yet 

 upon our possession of such data depends our being able to de- 

 termine in actual measures the diameter of the prisms into which 

 a given sort of basalt under given conditions shall split up. We 

 Phil May. S. 4. Vol. 50. No. 329. Aug. 1875. K 



