production of the Plasmatic Structure of Basalt. 209 



So far we have been regarding the mass of basalt as tabular 

 and horizontal, and cooling symmetrically from the top or bot- 

 tom surface or from both, in which cases the prisms are all 

 vertical and straight. We have yet to point out how differences 

 in the form, variations in thickness at different parts, and in 

 the positions of the principal cooling surfaces, can affect the 

 position in space of the axes of the prisms, and can cause 

 them to be no longer straight but curved in very various ways. 

 On the principles already developed, it is plain that the direction 

 of the axes of the prisms is determined by the form of the sur- 

 face passing through all the points in which splitting is actually 

 taking place, the separated sides of the prisms being always 

 normal to that surface. If, therefore, the mass be tabular and 

 horizontal, and cooled from the top surface only and uniformly 

 from it, the prisms will be vertical and straight, and will extend 

 from the top to the bottom of the mass, where they will rest upon 

 a thicker or thinner stratum of irregular fragments, as in fig. 10. 

 If, however, a like tabular mass be cooled from the top surface and 

 the bottom surface also, there will be formed two distinct ranges of 

 straight vertical prisms abutting upon each other somewhere be- 

 tween the top and the bottom of the mass, and separated by a stra- 

 tum of irregular angular fragments as in fig. 11 : the upper range 



nc.io 



F/C.li 



of prisms is produced by the top cooling causing splitting down- 

 wards, the lower range by the cooling from the bottom and 

 splitting upwards. The axes of any two prisms in the top and 

 bottom ranges need not coincide in position ; and the respective 

 lengths of the prisms in each range depends upon the rate of 

 cooling from the top and from the bottom respectively ; so that 

 if the cooling from each of these were the same, the separating 

 stratum of irregular fragments would be found in the middle of 

 the thickness of the mass. But in nature it almost always occurs, 

 for obvious reasons, that the rate of cooling from the top surface 

 greatly exceeds that from the bottom, and hence the upper 

 range of prisms is proportionately longer than the lower range. 

 Where the thickness of the mass is nearly level on the top and is 

 moulded to slight irregularities of the ground over which it has 

 Phil. Mag. S. 4. Vol. 50. No. 330. Sept. 1875. P 



