226 



ROBERT B. SOSMAN 



Benard 1 has recently shown photographically the identity of 

 pattern between his convection cells and the cross-section of the 

 basalt columns of the flow of Estreys (Haute-Loire), and has also 

 pointed out the qualitative differences between this pattern and 

 that produced by contraction. 



TABLE I 



Comparative Frequency of Polygons (Percentage) 



4. Frequency of angles. — The angles of convection columns 

 should approximate to 1 20 , while contraction columns will have a 

 large proportion near 90 . While the frequency of angles is a much 

 more logical criterion than the frequency of different polygons, it 

 is much more difficult to apply on account of the large number of 

 angular measurements to be made. Such a series was made with 

 great care by O'Reilly on the Giant's Causeway, and I have sum- 

 marized his results in Table II. O'Reilly's deduction from his 



TABLE II 

 Frequency of Angles in 206 Polygons of the Giant's Causeway 



Range No. of Occurrences Range No. of Occurrences 



to 75° 9 



64° 



75° 30' to 85 19* 



85° 30' to 95 S 6| 



95 30' to 105 103^ 



io5°3o' to 115 215 



ii5°3o' to 125 238^ 



i25°3o' to 135 236! 



i35°3o' to 145 143 



i45°3o' to 155° 18J 



i55°3o' to 175 5 



measurements was that the form of the columns had been governed 

 by the principal angles of the constituent minerals of the basalt, a 

 view which has not met with general acceptance. 



1 Compt. rend., CLVI (1913), 882-84. 



