8o TIMBER AND SOME OF ITS DISEASES, [chap. 



tracheides being about two mm. long, and only, say, 

 -^V nir^- 1^ diameter, the fluid in the longitudinal 

 cylinder (3) has only five barriers interposed for each 

 centimeter of length, whereas the tangential cylinder 

 (2) offers five hundred barriers for each centimeter 

 of length. 



Elfving then showed that if one uses thin plates 

 (one to two mm. thick) of the wood, the water passes 

 through as easily in the tangential direction as in the 

 longitudinal : whereas plates equally thin, but cut 

 in the plane of a tangent to the stem, will not allow 

 water to pass even under considerable pressures. 

 There is no mistaking the significance of the coin- 

 cidence that the water will pass so long as it meets 

 bordered pits, but will not pass in directions where it 

 meets none. 



Elfving recognised, however, that the partisans of 

 the imbibition theory might reply that these experi- 

 ments only demonstrated that the cell-walls transmit 

 differently in different directions, and that it was 

 necessary to test the alleged conductivity of the cell- 

 walls directly. 



He did this by the ingenious method of forcing 

 cacao-butter — which melts at 30° C, and does not 

 injure the walls — into the cut end of a piece of wood. 

 By colouring the cacao-butter with eosin, it was easy 



