138 OVERLYING AND TRAP ROCKS. 



They are not necessarily straight; being sometimes 

 carved, and even in more than one plane; while, 

 further, they are often found congregated together in 

 every mode of confusion, and also irregularly imbedded 

 in the amorphous rock, as in Staffa. Of one variation 

 in this form I have found but a single example, in 

 the island of Rum; the rock being disposed in im- 

 mense concretions, with convex surfaces of many 

 hundred yards in diameter, and cracked into polygons 

 resembling those of the septaria; the divisions being 

 wide at the surface, and contracting downwards, while 

 the polygons are, at the same time, horizontally 

 jointed, so as to separate in slabs. The whole appear- 

 ance resembles that which occurs in the superficial 

 consolidation of metals after fusion; the indurated 

 crust cracking in consequence of the greater expansion 

 of the fluid below. 



The columnar structure is not limited to one mem- 

 ber of this family, though vulgarly appropriated to 

 basalt ; since it occurs even in the antient porphyries, 

 and also in syenite, clinkstone, clay stone, augit rock, 

 and greenstone, as well as in basalt. Nor is it con- 

 fined to the masses; being often as decided and distinct 

 in the veins : the polygons being equally regular, and 

 presenting the same varying number of sides. These 

 are, more generally, placed at right angles to the 

 course of the vein, extending from one wall to the 

 other, as, very conspicuously, in the greater Curnbray, 

 but are also found parallel to it, and lying either ho- 

 rizontally, or vertically, as I have shown to be the 

 fact in Rum and in Cantyre. I need not now repeat 

 the probable causes of the columnar structure. 



The last peculiarity of configuration is the schistose, 

 formerly discussed. This occurs both in the masses 

 and in the veins ; being occasionally combined with the 

 columnar form, and taking place, then, either trans- 



