We observe that Fig. _d_, p. 27, from the want of proper shading, and error in not marking the dotted lines, does not serve well to illustrate our purpose. This figure is intended to represent a tree of a short thick stem, dividing into four branches, springing out regularly in the manner of a cross, nearly at right angles with the stem. These branches cut over about three or four feet out from the division, form each one wing of a knee, and the stem, quartered longitudinally through the heart, forms the other wing. It is of great advantage to have four branches rather than two or three, as the stem, divided into four, by being twice cut down the middle, forms the wings nearly square; whereas, when divided {389} into two, the halves are broad and flat, and a considerable loss of timber takes place; besides, the two branches afford a thicker wing than the flat half of the stem does when squared. When the tree separates into three branches, the stem does not saw out conveniently; and when divided, the cleft part is angular, and much loss of timber also takes place in the squaring. When the stem divides into four branches, each of these branches coincides in thickness with the quartered stem, and the knees are obtained equally thick throughout, without any loss of timber. The four branches, at six or eight feet above the division, may with a little attention be thrown into a rectangular bend, and thus give eight knees from each tree. Knees are generally required of about eight inches in diameter, and three and a half feet in length of wing; but when they are to be had thicker and longer, a foot or more in thickness, and from four to ten feet in length of wing, they are equally in request, suiting for high rising floors or heel-knees. The directions for forming larch roots into knees after the tree is grubbed, are also not very explicit. The stem of the tree is cut over nearly the same distance from the bulb as the length of the root spurs; this quartered through the heart (in the same manner as above), forms one wing of the knee, and the four spurs form the other wings. The same advantage results from having four regular root-spurs in larch, as in having four regular branches in oak: the two processes are quite similar, only the roots in the one case, and the branches in the other, form one wing of the knees. {390} We have given no directions for the bending of plank timber. In larch, the wind generally gives the slight necessary bend to a sufficient proportion; and in oak, the trees frequently grow a little bent of their own accord. A foot-note has been omitted, stating, that the plan of bending young trees, by tying them to an adjacent tree, intended to be soon removed, belongs, as we are informed, to Mr Loudon. We regret that our allusion to the lamented Mr Huskisson was printed off before we knew of his death. * * * * *