56 ECONOMIC WOODS OF THE UNITED STATES 



TiEMANN, H. D.: The Strength of Wood as Influenced by Moisture, Cir- 

 cular 108, U. S. Forest Service, 1907, p. 42. 



Johnson, J. B.: Timber Physics, Part II, Bui. 8, U. S. Div. of Forestry, 

 1893, pp. 22-24. 



SHRINKAGE, WARPING, AND CHECKING 



The volume of wood is maximum when the cell walls -are 

 saturated with water. When this condition exists the presence 

 or absence of free water in the cell cavities and the intercellular 

 spaces does not affect the volume. When the cell walls begin 

 to dry, they become thinner, but do not contract to an appreciable 

 extent longitudinally. A dry wood cell is therefore of practically 

 the same length as it was in a green or saturated condition, but 

 is cmaller in cross section, has thinner walls and a larger lumen. 

 According to Nageli's hypothesis, the cell wall is composed of 

 aggregations in crystalline form of minute parts or micelles. 

 These micellce are separated by films of water which become 

 thinner as the wall dries and thicker as it swells. This shrinkage 

 is roughly proportional to the thickness of the walls, and in con- 

 sequence the denser woods or the denser portions of a wood 

 shrink more than those less dense. ^ ' 



Inasmuch as wood is not a homogeneous substance, but an 

 intricate structure composed of cells exhibiting from moderate 

 to extreme variation in shape, size, thickness of walls, and more 

 especially in arrangement, it follows that shrinkage cannot be 

 uniform throughout any specimen. Late wood, being denser, 

 shrinks more than early wood. The ray cells, with their longest 

 diameters for the most part at right angles to the direction of the 

 other elements, oppose radial shrinkage and tend to produce 

 longitudinal shrinkage of wood. Only in the tangential direc- 

 tion are these otherwise opposing forces parallel. For this reason 

 as well as the fact that the denser bands of late wood are 

 tangentially continuous, while radially they are separated by 

 alternate zones of less dense early wood, wood usually shrinks 

 more than twice as much tangentially than it does radially. In 

 all cases, however, shrinkage parallel to the vertical axis is very 

 slight, oue-tenth to one-third of one per cent, and is maximum 

 in woods with curly or wavy grain or with large or very abundant 

 rays. 



The following table gives the results of a series of shrinkage 



