In fresh sap-wood, 60 per cent. of the water is in the cell cavities, 35 per cent. in the cell walls, and only 5 per cent. in the protoplasm. There is so much water in green wood that a sappy pole will soon sink when set afloat. The reason why there is much less water in heart-wood is because its cells are dead and inactive, and hence without sap and without protoplasm. There is only what saturates the cell walls. Even so, there is considerable water in heart-wood.[2] The lighter kinds have the most water in the sap-wood, thus sycamore has more than hickory. Curiously enough, a tree contains about as much water in winter as in summer. The water is held there, it is supposed, by capillary attraction, since the cells are inactive, so that at all times the water in wood keeps the cell walls distended. THE SHRINKAGE OF WOOD. When a tree is cut down, its water at once begins to evaporate. This process is called "seasoning."[A] In drying, the free water within the cells keeps the cell walls saturated; but when all the free water has been removed, the cell walls begin to yield up their moisture. Water will not flow out of wood unless it is forced out by heat, as when green wood is put on a fire. Ordinarily it evaporates slowly. [Footnote A: See _Handwork in Wood_, Chapter III.] The water evaporates faster from some kinds of wood than from other kinds, _e.g._, from white pine than from oak, from small pieces than from large, and from end grain than from a longitudinal section; and it also evaporates faster in high than in low temperatures. Evaporation affects wood in three respects, weight, strength, and size. The weight is reduced, the strength is increased, and shrinkage takes place. The reduction in weight and increase in strength, important as they are, are of less importance than the shrinkage, which often involves warping and other distortions. The water in wood affects its size by keeping the cell walls distended. If all the cells of a piece of wood were the same size, and had walls the same thickness, and all ran in the same direction, then the shrinkage would be uniform. But, as we have seen, the structure of wood is not homogeneous. Some cellular elements are large, some small, some have thick walls, some thin walls, some run longitudinally and some (the pith rays) run radially. The effects will be various in differently shaped pieces of wood but they can easily be accounted for if one bears in mind these three facts: (1) that the shrinkage is in the cell wall, and therefore (2) that the thick-walled cells shrink more than thin-walled cells and (3) that the cells do not shrink much, if any, lengthwise. (1) The shrinkage of wood takes place in the walls of the cells that compose it, that is, the cell walls become thinner, as indicated by the dotted lines in Fig. 35, which is a cross-section of a single cell. The diameter of the whole cell becomes less, and the opening, or lumen, of the cell becomes larger.