OBSERVATIONS ON BASALT, &C 179' 



matter muft: traverfe the greateft fpace, in order that the con- 

 traction may be performed without breach of continuity ; 

 therefore, if it be an extenfive tabular mafs, it will be divided 

 into prifms, by fiffures perpendicular to its furfaces. The 

 power of aggregation would determine thefe prifms to be 

 hexagonal, as that form contains the greatefi quantity of 

 matter in the leaft furface, of any prifms that can be united 

 without interpofing prifms of other forms. But this would 

 require the texture, the contraction, the thicknefs of the 

 mafs, and its adhefion to furrounding fubftances, to be 

 every where precifely the fame ; and, as thefe conditions 

 can never be fulfilled in an extenfive formation, all the irregu- 

 larities that are found muft neceffarily enfue. The fame rule 

 that determines the filfures of a tabular mafs to be perpendicular 

 to its furfaces, muft determine the rents in a fpheroid to be 

 directed from its periphery to its centre. 



Though thefe con fi derations may be fufficient to explain the Tney do not ac- 

 tendency to divifion into prifms, which is fo generally extend- "^fo/n^of"" 

 ed, and which has produced many of thofe abortions that bafaltes, 

 have been dignified with the name of columns, becaufe they 

 have occurred in lavas and in rocks of trap formation, they 

 are utterly inadequate to illuftrating the formation of the 

 more perfect bafaltic prifms : they offer no means of accounting 

 for the extreme regularity of the fides and the precifion of 

 the angles, for the articulations, for the clofe contact in which 

 the perfect columns are placed to one another, not for their 

 mutual adhefion, which is fo ftrong, that it often requires 

 confiderable violence to feparate them. Thefe fads are in 

 abfolute contradi&ion to all idea of retreat or contraction, and 

 feem to me to coincide perfectly with the explanation of 

 their origin which I have already prefumed to lay before 

 you. 



I have the honour to be, &c. 



GREGORY WATT, 



]S % Improved 



