IV.] VARIOUS THEORIES, &c 89 



No less important was the discovery that the air is 

 not at equal pressure throughout, but that it is less 

 dense in the upper parts of the tree than in the lower. 



Hartig's views, as then expressed, were as follows : 

 The water, enclosed together with air in a tracheide, 

 &c., is supported in the tubular cavity by capillarity, 

 so that its weight cannot make itself felt downwards 

 through the closing membranes of the bordered pits ; 

 in true vessels the individual water-columns are 

 suspended, separated by air-bubbles. When transpira- 

 tion is active, and the amount of water in the tree 

 tends to diminish, the air in the upper parts becomes 

 much more rarefied than that in the lower : for 

 instance, in the wood of the branches of the crow^n, 

 the air may be expanded to five times its original 

 volume, while that in the lower parts of the stem 

 expands simultaneously to only twice its volume. 



This diminution of pressure as we ascend must 

 exert a relatively powerful lifting force from tra- 

 cheide to tracheide, the greater pressure of the air 

 below driving the water through the membranes of 

 the bordered-pits. 



If the supply of water from below is arrested (as 

 was done in an old spruce by sawing to the depth of the 

 non-conducting inner wood) the density of the air 

 slowly becomes equalized throughout the whole 



