416 PROCEEDINGS OF THE GEOLOGICAL SOCIETY. [JunC 16, 



tabular mass above, whilst the larger vein terminates abruptly in a. 

 rounded knoll. 



Fig. 2. — Trap-veins in the Cliffs near Losset. 



a. Sandstone and limestone. 



b, b. Trap-veins. 



Further north similar veins traverse the massive or bedded trap, 

 and often project from its surface like rude walls. They are well 

 exhibited on many parts of the terrace or raised beach, having appa- 

 rently resisted the denuding influence of the sea better than the hori- 

 zontal mass. They are also, I think, often of a harder and more 

 crystalline structure than the latter, from which they must differ in 

 point of age. 



One of the most interesting exhibitions of the trap-rocks occurs on 

 the east coast, near the ruins of Kilhousland church, to the north of 

 Campbeltown harbour. In this place a vein of dark ferruginous 

 greenstone, divided at right angles to the sides into nearly horizontal 

 columns, intersects a mass of light-grey trap which forms the shore 

 for a considerable distance. In some places the latter becomes highly 

 concretionary, and is divided into vertical columns which cover the 

 shore like a miniature Giant's Causeway. These columns are sepa- 

 rated from each other by veins of calc-spar, haematite, and green 

 earthy carbonate of copper, and the whole rock is in some places im- 

 pregnated with the latter mineral. The haematite is either compact 

 and siliceous, with portions of red hornstone, or it is more granular, 

 with veins and nodules of calc-spar and grains of green earth, or con- 

 tains druses of calc-spar lined with crystals of quartz. The haematite 

 veins are sometimes 3 or 4 inches wide, but generally much less. 

 Occasionally the included column has been entirely removed, when 

 the intercolumnar veins remain like empty honey-combs. 



The structure of this singular columnar mass might give rise to 

 many speculations. The usual theory of the formation of basaltic 

 columns is that they have been formed by the contraction of the mass 

 in cooling, and their position, always at right angles to the walls or 

 cooling surface, so far confirms this theory. In general the prisms 

 fit each other very closely, with no wider spaces than the contraction 

 of the mass might be supposed capable of producing. Here, however, 

 we find columns of about a foot in diameter separated by intervals of 



