for wheels, segments, etc. The table is made adjustable for cutting bevels or draught.

The circular saw, as used in pattern making, is mainly applied to roughing-out purposes. More stuff being left for finishing by hand, than would be the case were the work sawn with a band or a jig saw. The circular is a very useful saw, however, especially for roughing out rabbets and similar work. The table for pattern maker's use

should adjust to saw at a bevel, and should rise and fall adjustably at one end, so that the saw may project more or less above the surface of the table; the height of the top of the saw, from the table surface vertically beneath it, regulating the depth of the groove the saw will cut. The same saw being used for slitting and cross-cutting purposes, the teeth are filed slightly pointed, and thus answer both purposes. The circular saw, shown in Fig. 242, is designed for pattern maker's use; the gage being operated by a screw operated by the handle shown.

Roller feeding planing machines are not properly adapted for pattern making, because the pressure of the roller springs the work out of true. They may, it is true, be used upon work too thin to be held in other planers, but there being in any case no assurance of truth (that great desideratum for pattern makers) their employment, even for

The

thin work, is not advisable. A traverse planing machine, however, is a very useful tool, especially upon segment work; hence such a machine is shown in Fig. 243. frame is boxed, to secure rigidity with compactness and lightness. The feed is a hand one, as is preferred by all pattern makers, because it admits of rapid manipulation. The framing being open at the front, gives easy access to the cutters, and admits work of greater width.

Fig. 244

In every pattern shop hot glue is a primary necessity, and steam is by all means the best medium of keeping the same heated ready for use. In Fig. 244 is shown a steam glue heater; the outer casing containing the water, there being a glue pot on each side of the upper face, and a pot for hot water in the center. In the absence of steam, the ordinary glue pot, heated by gas or a spirit lamp, is employed.

The glue used in pattern making must be of the best quality, well boiled, and applied properly hot; because, notwithstanding the varnish, patterns are affected by the moisture of the molding-sand, and from rapping the pattern

to loosen it in the mold. Defective gluing - or in fact any but the best executed gluing will rapidly show itself, and impair the value of the pattern.

CHAPTER XVIII.

SHRINKAGE IN CASTINGS.

To allow for the shrinkage in castings, the pattern is not infrequently made in form and size to meet the requirements of any known case. Suppose, for example, that the surface of a large casting is found to be hollow, then that surface upon the pattern will be made sufficiently rounding to allow for the shrinkage, thus giving a casting with the desired flat surface. In large bodies of metal the shrinkage is always sufficient to demand an allowance therefor by the pattern maker; and it always takes place in the largest mass of the metal. The directions of this shrinkage are thus given for particular forms by Mr. ALFRED C. WATKINS:

SOLID CYLINDERS.

In the case of a shaft, or other solid cylinder, it will be noticed that the surface of the casting at the ends will be slightly depressed. This is occasioned by the surface of the cylinder being cooled by the walls of the mold first, and setting, while the central portion yet remains fluid or soft. In a few moments more the central portion cools, and in shrinking draws in the ends of the cylinder, the outer crust acting as a prop or stay to the atoms of metal adjacent to it. If this theory be correct, the depression should take the form of an inverted cone, owing to the gradual checking of the shrinkage as it approaches the outer crust. In practice this will be found the case-the obtuseness of the angle being greater or less, according to the nature of the iron to shrink.

GLOBES.

In the case of solid globular castings, the heart or central point within will usually be found hollow or porous, owing to the following causes: The walls of the mold cooling off the outer surface, causes it to set immediately; the interior, cooling from the exterior inward, endeavors to shrink away from the outer crust, which resists its so doing; hence, the interior is kept to a greater diameter than is natural, and there being but so much metal in the entire mass, the atoms are drawn away from the central point toward all directions, to supply the demand made by the metal in shrinking.

DISKS.

In the case of flat round disks or plates, they will usually be found hollow on the top side, although in some cases the hollow is on the bottom side. This is owing to the following causes: The top and bottom faces, together with the outside edge, become set first through contact with the mold, leaving the center yet soft. When the center shrinks a severe strain is put on the plate by an effort to reduce its diameter, which the outer edge resists. Now, if the cop be thin, the heat will radiate rapidly in that direction, causing the outer or top sido to set first; the under side, setting later, will drag the top side over with it, causing it to round up on top and dish in the bottom. Or if the pattern be not perfectly true in every direction, the strains first spoken of will cause any curved portion to become more exaggerated. If the pattern be perfectly true, cop and drag of the same thickness, and both rammed evenly, there is no reason why the plate should not come out perfectly true, the strains being all self-contained in the same plane and balanced. If the plate, however, have an ogce molding projecting downward around the edge, it will likely be depressed on the top surface when cast. This is due to all the surfaccs being set alike and at the same instant, excepting the metal within the corners, which, containing the most metal in a mass, will shrink last of all. When this does shrink, its tendency is to pull over the top side of the molding toward the plate, which being soft, although set, will be forced downward at the edges, giving a chance for the strains within the plate, as above described, to aid in the distortion.

ROUND AND SQUARE BARS.

These strains are similar in both, and are already treated of under solid cylinders. There is another feature, not before spoken of, which is rather curious. If two bars of the same dimensions and mixture of iron be heated to the same temperature, the one allowed to cool in the mold, the other plunged while hot into water, the latter will be found to have shrunk the most. This is due to the particles about the surface having been enabled, by the softness of the interior metal, to get closer to each other than they could have done if the material had cooled slowly.

RECTANGULAR TUBES.

These are usually cast with a core, which has a tendency to retain the shape of the casting; still the flat sides will show a tendency to bulge up slightly at the middle. This is due to much of the same