PRISMATIC STRUCTURE IN IGNEOUS ROCKS 219 



are those produced by contraction, not by convection (see p. 225). 

 A photograph of the occurrence is shown in Fig. 1 . An examination 

 by Wherry of the cross-section of one of the small "dikes" shows 

 that it has an irregular boundary, that it grades off without a sharp 

 break into the surrounding rock, and that it is more coarsely crys- 

 talline than the surrounding material. It appears to be, therefore, 

 a case of prismatic structure due to contraction in physically hetero- 

 geneous material, and quite distinct from the usual type of con- 

 traction prisms. Dr. N. L. Bowen, of this laboratory, informs me 

 that he has seen a similar structure in the upper surface of a diabase 

 sill north of Lake Superior. 1 



CONVECTION HYPOTHESIS 



E. H. Weber 2 described in 1855 a phenomenon observed by him 

 on microscope slides on which a solid was being precipitated from 

 alcohol-water mixtures. The liquid was observed to circulate and to 

 divide itself up into regular polyhedral cells. A similar phenomenon 

 was observed by James Thomson 3 in 1882, in a soap solution. It 

 remained for the French physicist Benard, 4 in 1900, to make a 

 really thorough study of the subject, and his experiments have 

 brought out a number of new and interesting facts. 



A polygonal structure is easily produced in a layer of liquid 

 which is shallow in comparison with its horizontal extent, and which 

 is losing heat from its upper surface or is gaining heat from its 

 lower surface. If the top surface is cooler than the bottom, then 

 the colder and denser liquid at the top tends to sink and the warmer 

 bottom layer to rise, and convection currents must be set up. If 

 the conditions are uniform and constant, a steady state of flow of 

 some kind must ultimately be set up. In a flat liquid sheet of 

 indefinite extent this state of flow must take the form of parallel 

 rising and descending currents, and these will flow with minimum 



Canada, Bur. Mines, Ann. Rep., XX (1911), 125-26. 



2 Pogg. Ann., XCIV, (1855) 45 2 -59- 



3 Phil. Soc. Glasgow, Proc. XIII, (1882), 464-68. Thomson recognized the simi- 

 larity of the pattern to that of the Giant's Causeway. 



t H. Benard, Les tourbillons cellulaires dans une nappe liquide, etc., thesis, Paris, 

 1901; Rev. gen. ScL, XI (1900), 1261-71, 1309-38- 



