water from the wood, and is found to be 1·56, compared with an equal volume of water taken as unity. It is the varying quantities of this wood substance, and of air and water in the cavities, which make the density of different pieces of oak vary so much.
(3) The proportion of sap contained in the cavities of the vessels, cells, etc., of course differs at different times. In the spring, just as the buds are opening, the quantity of water increases more and more up to about July, when the maximum is attained; the proportion of water to solids then sinks until October, when the leaves fall; it increases again up to Christmas-tide, and then sinks to the minimum in the coldest part of the winter. The proportion of water to the total weight of the felled wood may vary from 23 to 39 per cent.(4) Obviously the loss of water on drying causes shrinkage of the wood, and although oak shrinks very little in the direction of its length (0·028 to 0·435 per cent), the effect is very marked in other directions. In the radial direction—i.e., in the direction of the medullary rays—it may shrink from 1 to 7·5 per cent of its measurement when first felled; and in the direction vertical to this—i.e., parallel to a tangent to the cylindrical stem—the variation is from 0·8 to 10·6 per cent. Of course, green oak shrinks much more than seasoned and older wood, the process of seasoning being, in point of fact, the period of chief shrinkage. It is said that wood from the variety sessiliflora shrinks more than that of the variety pedunculata, but it may be doubted