Crystalline Structures of Rocks. 65 



cause may have produced the columnar form in the trap rocks. 

 But the assertion is unfounded ; inasmuch as columnar lavas are 

 found where no water can have been present, and amorphous ones 

 occur beneath the sea. 



The next case which requires notice in this general account of 

 the columnar structure, is where that form occurs in sandstone ; 

 the only two instances with which I am acquainted, are found in 

 the island of Rum, and at Dunbar in Scotland. 



The columns that occur in the sandstone of Rum are of small di- 

 mensions, not exceeding a few inches in diameter. They lie in the 

 stratum in perfect contact, presenting the usual intermixture of 

 polygonal forms ; and what is especially necessaiy to notice, they 

 are covered by a mass of basalt. At Dunbar, the sandstone in 

 which the columnar arrangement is found, is that which is known 

 to be the lowest of the secondary strata, and which, throughout a 

 great extent of country, presents only the usual stratified character. 

 The columns are limited to a small space, but are of considerable 

 dimensions ; attaining to two feet or more in diameter, and to a 

 length of 15 feet or upwards. Where this columnar structure 

 occurs, the character of the rock is changed ; becoming more com- 

 pact, harder, and in some places, passing into a perfect but coarse 

 jasper. In addition to this, it presents the indications of an internal 

 concretionary structure; similar to that which might be inferred to 

 exist in the columns of trap, from the mode in which they are 

 found to desquamate. The transverse sections of each prism are 

 marked by concentric lines of different colours, whitish and reddish; 

 which conform accurately to the sides and angles towards the ex- 

 terior, but become gradually curved as they approach the centre ; 

 indicating the probable existence of a spheroidal nucleus. This dis- 

 position is unconnected with any agency of the atmosphere. 



As it appears to me that this example, and that of the columnar 

 shales, or argillaceous iron-stones, as they have been called, are in 

 every respect analogous and admit of the same reasoning, it will 

 be as well to describe such examples of these as may be useful in 

 the arguments to be deduced from them. This modification occurs 

 on the large scale in Arran ; the prisms being of large diameters, 



Vol. XVIII. F 



