A pith ray shrinks in height and width, that is, vertically and tangentially as applied to the position in a standing tree, but very little in length or radially. The other elements of the wood shrink radially and tangentially, but almost none lengthwise or vertically as applied to the tree. Here, then, we find the shrinkage of the rays tending to shorten a stick of wood, while the other cells resist it, and the tendency of a stick to get smaller in circumference is resisted by the endwise reaction or thrust of the rays. Only in a tangential direction, or around the stick in direction of the annual rings of growth, do the two forces coincide. Another factor to the same end is that the denser bands of late wood are continuous in a tangential direction, while radially they are separated by alternate zones of less dense early wood. Consequently the shrinkage along the rings (tangential) is fully twice as much as toward the centre (radial). (See Table XIV.) This explains why some cracks open more and more as drying advances. (See Fig. 27.) Although actual shrinkage in length is small, nevertheless the tendency of the rays to shorten a stick produces strains which are responsible for some of the splitting open of ties, posts, and sawed timbers with box heart. At the very centre of a tree the wood is light and weak, while farther out it becomes denser and stronger. Longitudinal shrinkage is accordingly least at the centre and greater toward the outside, tending to become greatest in the sapwood. When a round or a box-heart timber dries fast it splits radially, and as drying continues the cleft widens partly on account of the greater tangential shrinkage and also because the greater contraction of the outer fibres warps the sections apart. If a small hardwood stem is split while green for a short distance at the end and placed where it can dry out rapidly, the sections will become bow-shaped with the concave sides out. These various facts, taken together, explain why, for example, an oak tie, pole, or log may split open its entire length if drying proceeds rapidly and far enough. Initial stresses in the living trees produce a similar effect when the log is sawn into boards. This is especially so in _Eucalyptus globulus_ and to a less extent with any rapidly grown wood. The use of S-shaped thin steel clamps to prevent large checks and splits is now a common practice in this country with crossties and poles as it has been for a long time in European countries. These devices are driven into the butts of the timbers so as to cross incipient checks and prevent their widening. In place of the regular S-hook another of crimped iron has been devised. (See Fig. 28.) Thin straps of iron with one tapered edge are run between intermeshing cogs and crimped, after which they may be cut off any length desired. The time for driving S-irons of either form is when the cracks first appear.