466 EMERSON— RECURRENT TETRAHEDRAL DEFORAIATIONS. 



lines, which might be sometimes cemented so as to be Hnes of great- 

 est strength. It would not need to fold at the same places in suc- 

 cessive compression periods. The other points where three lines 

 join on the fulcrum line are wholly unlike the corresponding points 

 on a trap column. They are indeed points where three almost non- 

 existent lines meet, since tension and motion die out as the fulcrum 

 line is reached. During the subsequent compression period also 

 these points are places of minimum compression and so of minimum 

 elevation, but they are the points where the greatest protuberances — 

 the continental shields — must be. 



It is, moreover, hard to see how the three polar fissures can 

 exert any influence across this dead space to locate the corresponding 

 fissures which stretch across the equator, since the maximum ten- 

 sion by which these fissures are formed is at the far distant equator 

 where it would be more probably relieved by fission along three lines 

 at 120° (after the manner of trap), radiating from centers on the 

 equator and at convenient distances apart, rather than by lines or 

 bands slanting across the equator 8,000 miles apart. 



I have seen where the triassic sandstone has been stripped off 

 the trap and found no elevation at the junction edges of the surface 

 of the columns or depressions at their centers and the same is true 

 of mud cracks. There is rather a slight depression where the col- 

 umns join. The ball and socket structure is a deep-seated one, and 

 the ridges along the edges of adjacent columns and the elevations at 

 the corners are not upthrusts in any sense. The six-sided column 

 has first formed by shrinkage and rupture, and no further action 

 takes place across the ruptured surfaces, then later shrinking and 

 consequent fissuring inside each column separately have produced a 

 " spheroidal parting " inside each individual column and it is this 

 curved parting which forms the apparent hollow when the column 

 falls in pieces, or when several columns have been eroded to a com- 

 mon level forms the adjacent hollows bounded by the intervening 

 ridges and corner projections. There is no trace of a longitudinal 

 motion of the central part of the basalt column up or down or side- 

 wise. Indeed the blocks into which the column breaks will be con- 

 cave upward for a while and then be followed by a double concave 

 block and then will be convex upward for a time and then be fol- 



