HL, &c. equal to B1, C2, D3, &c.; and the ordinates of the inner curve equal to CM, DN, &c.; then, through the points E, I, K, L, &c., draw a curve; and within, through the other points, draw another curve; and this will form the face-mould for the sash-head.

Figure 3 is the development of the soffit of the under side of the sash-head. Figure 4, that which is applied to the outer or convex side of the same: so that, by cutting away the superfluous wood on the outside of the space contained by the two lines, the sash-head will fit the surface of a cylinder made to the radius of the plan of the window.*

91. To find the Radial Bars.-Let ny, fig. 1, be the place of a radial bar. In ny take any number of points, n o p q, &c., and draw the perpendiculars nv, ow, px, qy, &c., to ny. Draw also, ni, ok, pl, qm, &c., perpendicular to the springing line of the elevation, cutting that line at the points a, b, c, d, &c.; and cutting the plan of the convex side of the sash at the points e, f, g, h, and the concave side at i, k, l, m, &c. Make the distances nv, ow, px, qy, &c., respectively equal to ai, bk, cl, dm, &c.; and, through the points v, w, x, y, &c., draw a curve, which will form the convex edge of the radial bar. In the lines nv, ow, px, qy, &c., make the distances nr, os, pt, qu, &c., each respectively equal to ae, bf, cg, dh, &c.; and, through the points r, s, t, u, &c., draw another curve, which will give the inner edge of the radial bar.

OF THE CONSTRUCTION OF WINDOW SHUTTERS.

92. WHEN the walls of a room are sufficiently thick, the window-shutters are made to fold into a recess within the thickness of the wall, and for the more free admission of light and air the window recess has its sides splayed to a greater or less angle, the splay most frequently adopted appears to be from one-third to two-fifths of the depth of the window recess.

In plate LIX we have represented the parts of a SASH-WINDOW, and its SHUTTERS, finished in this manner; and in all its detail according to the style of the working drawings of the most eminent architects of the present time.

Figure 1 is a vertical section through the window, with parts removed to enable us to give it on a larger scale on the plate.

Figure 2 is a plan of the same, showing the shutters as folded back into the boxings; for so the recess formed behind the architrave to receive them is called. Fig. 3 shows the elevation of the lower part of the window and shutters, and the construction of the plain back and plinth to the window recess.

In fig. 1 S is the sill of the sash-frame; W the stone sill with a sinking to prevent water collecting at the edge of the wooden sill so as to decay it; and B is the bottom rail of the sash. M is the meeting bars of the sashes; T the top rail of the upper sash; and H is the head of the sash-frame, with the soffit grooved into it. A is the architrave, and P its plinth.

In fig. 2, F is the pulley stile of the sash-frame. We recommend this mode of rebating for the beads into the pulley stiles from many years experience of its utility; and we also recommend fixing the beads by means of screws with small pieces of brass let in and counter-sunk for their heads.

The methods in this and the preceding article, (Arts. 89 and 90,) apply only to the case when the bars and stiles are parallel to the radius of curvature at the middle of the breadth; but the bars and the stiles ought to be perpendicular to the curve, and then the true form for the mould may be found from the development of part of the surface of a wedge formed solid, having a semi-circular end, on the principles we have laid down in Carpentry, Art. 50 and 53.

METHODS OF TAKING DIMENSIONS AND SETTING OUT WORK.

89

forms methods of his own, and, merely from his being most familiar with his own processes, he will, following his own methods, do his work in a better manner than by strange ones, even if to an unprejudiced mind the methods he followed were evidently inferior.

The end and aim of a joiner, in all these operations, is to avoid the peculiar imperfections and disadvantages of his materials, and to do this at the least expense of time and wood. The straightness of the fibres of wood renders it unfit for curved surfaces, at least, when the curvature is considerable. Hence, short pieces are glued together as nearly in the form desired as can be, and the apparent surfaces are covered with thin veneers; or the work is glued up in pieces that are thin enough to bend to the required form. Sometimes a thin piece of wood is bent to the proposed form, upon a saddle or solid cylinder; and blocks are jointed, and glued upon the back; the whole is then allowed to become completely dry, and it will preserve the form that has been given to it by the saddle. But, when a piece of work is glued up in parts, it should be extremely well done; and, not as frequently happens with the joints black and irregular, and springing open in places. The difficulty of doing work of this kind well, has led to the trial of a variety of other methods.

48. If a piece of wood be boiled in water for a certain time, then taken out, and bent while hot into the proposed form, and it be retained in that form by screws or wedges till it be perfectly dry, it is found to preserve nearly the same figure that has been given to it. The quantity it springs back, when relieved, is not easily allowed for; and it is equally uncertain how long it may continue to return towards its natural straightness.

49. The same effect may be produced by steaming, as by boiling wood; and, indeed, more < effectively. Both methods have been long practised, to a considerable extent, in the art of ship-building; but we are not aware that any general principles have yet been discovered either by experiment or otherwise, that would enable us to apply it with that degree of certainty and precision which is required in joinery. It has frequently been tried to bend wood by these means for the joiner's purpose, but we still want to know what relation there is between the curve to which it may be bent, and that which it will retain; and also the degree of bending we may with safety give to a piece of wood of a given thickness; for it clearly must not be bent so as to injure the grain of the wood.

The time that a piece of wood should be boiled or steamed, in order that it may be in the best state for bending, should also be made the subject of inquiry by experiment; and this being determined, the relation between the time and the thickness of the pieces should be ascertained.

For the purposes of joinery, we think that the process might be improved by saturating the pores of the convex side of each piece with a strong solution of clear glue immediately after bending it; for, by filling in this manner the extended pores, and allowing the glue to harden thoroughly before relieving the pieces, they would retain their shape better.

METHODS OF TAKING DIMENSIONS AND SETTING OUT WORK.

50. TAKING dimensions of plain square work is so simple, that we need not consider it; but, when irregular figures are to be framed, it requires more skill. The methods to be followed, in such cases, depend upon the method of describing a triangle, when its three sides are given. Thus, let ABC (fig. 6, pl. XLVII,) be the three given sides of a triangle; in any convenient