FACTORS IN COLUMNAR STRUCTURE IN LAVAS 



469 



between 120° and 130°. From these observations it is clear that it 

 is not safe to classify columns as due to contraction or convection 

 by their shape and angles. 



Close to the "Sydenham Organ Pipes" there are a number of 

 huge squat irregular basaltic columns about 10' in diameter. The 

 flow is thin and the shape of the 

 great blocks seems to be deter- 

 mined by convection which left 

 its record in them by vesicles 

 and when the rock contracted 

 through loss of heat the cracks 

 formed along the lines of 

 vesicles (Fig. 16). The 

 vesicles are pulled round in 

 just the way one would expect 

 if convection currents were the 

 cause. There are centers 

 round which the currents 

 traveled and these tend to 

 be hollowed out leaving holes 

 3' to 4' across. It is possible 

 that the drawn-out vesicles arc- 

 due to the movement of the 

 lava before it came to rest, 

 but the fact that most of these 

 squat columns seem to have the 

 hollow or nucleus suggests con- 

 vection. Contraction due to 

 shrinkage subsequently tends 

 to crack the rock along the 

 lines where the currents rose 

 or sank. 



Convection columns in this area are either very unimportant or 

 absent so that the conclusions drawn in R. B. Sosman's paper in 

 the Journal of Geology, 191 6, do not find confirmation in the area 

 above described. 



Fig. 16. — Convection currents in basalt. 

 The direction of the vesicles and the 

 occurrence of the nuclei suggest that the 

 molten lava was disturbed by convection 

 currents. The vertical joints occur where 

 the long axes of the vesicles are vertical. 



