WATER TN THE CELL-CAVITIES. 



239 



This supposition, moreover, is proved to be correct by two further facts. If a piece 

 of fresh wood is placed in the fire, bubbles of gas are seen to be emitted with 

 violence, evidently because the air contained in the wood is expanded by the 

 heat and forced out at the transverse section, at the same time driving out 

 with it a portion of the water contained in the wood. Again, we find that a 

 piece of wood cut from a living tree, when laid upon water, gradually becomes 

 heavier and sinks deeper and deeper, because it absorbs water: this would be 



Fig. 200. — Tang^cntial longitudinal section tiirough the secondary wood of Ailanthris^iandjilosa. g g vessels ; 

 j^ medullary rays cut across;/ wood parenchyma ; / tracheVdes; 1/ librifonn cells. 



quite impossible if the cavities of the wood were actually already completely filled 

 with water. 



These observations prove then most decisively that just at the time of the most 

 rapid ascent of water in the wood, the cavities in the wood-cells cannot be filled with 

 water ; that, rather, cavities containing air and vapour must exist in them. By means 

 of more exact considerations and experimental researches, I have succeeded in finding 

 a method by which we are able to determine in any given piece of wood, by a 

 simple calculation, how large the empty space in the wood-cells is. For this purpose 

 the specific gravity of the wood cell-walls must first be exactly determined. Among 



