214 



UNSTRATIFIED OR IGNEOUS ROCKS. 



presents the appearance of a long pile of cord- wood. In some cases the 

 columns are curved and twisted in a manner not easy to explain ; some- 

 times, instead of columnar, a ball-stTiictuve is observed. 



Columnar Dike, Lake Superior (after Owen). 



Cause of Columnar Structure.— There is little doubt that this 

 structure is produced by contraction in the act of cooling. Many sub- 

 stances break in a prismatic way in contracting. Masses of wet starch, 

 or very fine mud exposed to the sun, crack in this way. In basalt the 

 structure is more regular than in any other known substance. The 

 subject of the cause of jointed columnar structure has been very ably 

 discussed by Mr. Mallet.* 



Volcanic Conglomerate and Breccia. — If a stream of fused rock, 

 whether from a crater or a fissure, run down a stream-bed, it gathers 

 up the 'pebbles in its course, and after solidification forms a conglom- 

 erate which differs from a true conglomerate (p. 171) in the fact that 

 the uniting paste is igneous instead of sedimentary. In a similar man- 

 ner volcanic breccias are formed by the flowing of a lava-stream over a 

 surface covered with rubble. 



The disintegration of volcanic rocks, and their transportation and 

 deposit, will of course give rise to aqueous conglomerates and breccias 

 composed of volcanic materials, which often are difficult to distinguish 

 from true volcanic conglomerates and breccias. These aqueous con- 

 glomerates and breccias of volcanic material pass by insensible grada- 

 tions into tufas, which, as already explained (p. 91), consist of fine vol- 

 canic material cemented into an earthy mass and often sorted by water. 



* Philosophical Magazine, August and September, 18*75. 



