208 Mr. R. Mallet on the Origin and Mechanism of 



breaking the prisms transversely, or pulling asunder lengthwise 

 or breaking transversely such prisms as may be unequally gripped 

 at different points by those surrounding them — these strains ori- 

 ginating in the later stages of cooling of the mass viewed as a 

 whole. On the other hand, where planes of easiest fracture are 

 more or less developed in the mass of unsplit-up basalt, their 

 effect will be to shorten the length between the joints, so that 

 they shall be found to occur within distances of each other less 

 than the diameter of the prism. This is frequently observed in 

 the Antrim and Scottish basalts ; and wherever that is the case, 

 so far as the writer has observed, the separate-joint blocks break 

 under the hammer more easily in the direction transverse to the 

 length of the prism than in any other. We have yet to consider 

 some of the circumstances which render the prisms into which 

 the mass splits more or less irregular as to the lengths and the 

 number of their sides. As has been shown, in homogeneous and 

 isotropic basalt cooling symmetrically, the prisms can only be 

 hexagonal and equilateral. Irregular prisms, however, are to 

 be found in most basaltic regions, having from three to seven or 

 even more sides, and often far from equilateral. Without going 

 into prolix detail, it will be easily seen that differences of mole- 

 cular structure in different directions in the yet unsplit mass, the 

 unequal dispersion throughout of cavities or heterogeneous im- 

 bedded materials, differences of conductivity in nearly contiguous 

 parts of the mass, slight differences in the contraction in differ- 

 ent directions and the consequent greater width of the fissures 

 separating the prisms in one direction rather than in another, 

 and slight differences in the rate of cooling of adjacent parts of 

 the mass by inequalities in the contour of the cooling surface or 

 in the "jacketing" of detrital matter which may cover it, 

 or in the movements of superincumbent water, all tend respec- 

 tively to perturb the regularity of the hexagonal prism, which, 

 however, is never observed to disappear as the ruling or typical 

 form of the basaltic prism. Walking over the tops of a large 

 pavement of basaltic columns such as those exposed at the 

 Giant's Causeway and elsewhere, one sees places in which the 

 hexagonal form becomes irregular, and gives place, though for a 

 very small extent of surface, to quadrangular and other irregular 

 forms, sometimes even triangular or wedge-shaped ; but passing 

 a short distance we find these rapidly returning to the normal 

 hexagonal type through intervening prismatic forms, which usu- 

 ally plainly indicate that, whatever may have been the origina- 

 ting cause of the disturbance of form, the tendency is to return 

 to the hexagonal prism through such intermediate forms of irre- 

 gular prisms as require the least amount of molecular work to 

 produce their separation, the principle of least action thus being 

 ever and pervadingly present. 



