The construction of patterns for iron castings, requires not only the observance of all the particulars conveyed on pages

327 to 329, but in addition, the large size of the models, the peculiar methods employed in moulding them, and the nearly' inflexible nature of the iron castings when produced, call for some other and important considerations; and which should not be entirely overlooked, even in works of comparatively small size, or it may lead to failure and disappointment.

Thus, it becomes necessary to make patterns in some degree larger than the intended iron castings, to allow for their contraction in cooling, which equals from about the ninety-fifth to the ninety-eighth part of their length, or nearly one per cent. This allowance is very easily and correctly managed by the employment of a contraction rule, which is made like a surveyor's rod, but one-eighth of an inch longer in every foot than ordinary standard measure. By the employment of such contraction rules, every measurement of the pattern is made proportionally larger without any trouble of calculation.

When a wood pattern is made, from which an iron pattern is to be cast, the latter being intended to serve as the permanent foundry pattern; as there are two shrinkages to allow for, a double contraction rule is employed, or one the length of which is one quarter of an inch in excess in every foot. These rules are particularly important in setting out alterations in, or additions to, existing machinery; the latter is measured with the common rule, and the new patterns are set out to the same nominal measures, with a single or double contraction rule as the case may be, the three being made in some respects dissimilar to avoid confusion in their use; the entire neglect of contraction rules, incurs additional trouble and uncertainty.*

Patterns for iron castings are much more frequently divided into several parts than those for brass; for instance, the division into two equal parts after the manner of fig. 175, p. 339, (but without reference to the under-cutting,) is very common, as both the pattern and flask separate when the top part is lifted, and the halves of the pattern can then be drawn out from the halves of the flask with much less risk of tearing down the sand.

Referring to p. 319, fig. 141, if small, would be moulded as represented with false cores or drawbacks; but if it were a large fluted column, the iron founder would employ a solid two-part flask; the shaded parts would together represent the body of sand in the drag, and the pattern would be made in three parts something like a boot-tree. When the top flask had been lifted, the central slice of the pattern, extending from the two upper to the two lower angles would be withdrawn vertically, and the two outer pieces would be released sideways. The general rule is to divide the circumference of the pattern into six equal parts, and to let the central slice equal one of them in width.

The figures, 189 and 191, representing two parts of a slide-rest, and the pedestal, 128, are some amongst many of the common examples of the division of the patterns; and with which may be associated, the numerous subdivisions of the mould instead of the pattern by the employment of cores, many applications of which have been also explained. All these matters display much interesting and ingenious contrivance, resorted to either to render possible the operation of moulding, or to facilitate its performance.

* The contraction of brass is nearly three sixteenths of an inch in every foot, but from the small size of bras3 castings, the contraction rule is less required for them as the differences may be easily allowed for without it.

Iron castings weigh about fourteen times as much as the ordinary deal and fir patterns from which they are made, that being nearly the ratio of the specific gravities of those materials. All these matters are entered into in the author's pamphlet, "A New System of Scales of Equal Parts," and his paper "On a scale of Geometrical Equivalents for Engineering and other Purposes," Lond. and Edin. Philos. Mag., July, 1838, wherein are described numerous applications of scales of equal parts to the purposes of drawing and calculation, and to the comparison and conversion of all kinds of measures, weights and other quantities.

To lessen the distortion of castings from their unequal contraction in cooling, it is important that the models should be nearly symmetrical. For example, bars or rods of all the sections in fig. 140, p. 319, may be expected to remain straight; perhaps g is the most uncertain, but if the lower fins of e and h were removed, their flat surfaces then exposed to the sand, would become rounding or convex in length, from the contraction of the upper rib being unopposed by that of a similar piece on the under side. Bars and beams, the sections of which resemble the letter I arc of the most favourable kind for general permanence, and also for strength, and large panels may be cut out from their central plates to diminish their weight without materially reducing their stability. They are much used, not only in building, but also in the framing of machinery, which is in a great measure based upon the same general rules.

It is also of great importance, especially in castings of large size, that the thickness of the metal should be nearly alike throughout, so that it may cool at all parts in about the same time. Should it happen that one part is set or rigid whilst another is semi-fluid, or in the act of crystallising, there is great risk of the one part being altogether torn from the other and producing fracture. Or should the disturbing force be insufficient to break the casting, it may strain the metal nearly to its limit of tenacity or elasticity; so that a force far below that which the casting should properly bear, may break it in pieces.

An example of this is seen in wheels with very light arms and heavy rims or bosses. The arms sometimes cool so quickly as to tear themselves away from the still hot rim or nave; or when the arms are solidified without fracture, the contraction of the rim may so compress the spokes endways, as to dish the wheel, (in the manner of an ordinary carriage wheel,) and thereby strain the casting nearly or quite to the point of fracture. The arms are sometimes curved like the letter S, instead of being straight and radial; the contraction then increases their curvature with less risk of accident than to straight arms.* A more elegant way of avoiding the mischief was invented by Mr. Isaac Dodd, of the Horsley Iron Works, by placing the spokes as tangents to the central boss, in which case the contraction of the rim makes a small angular change of position in the boss; for the rim in thrusting the spokes inwards, causes the boss to twist round a little way with far less risk of fracture.*

* It appears to be often desirable to supersede the straight diagonal braces of iron cantings, by curved lines, which are both more ornamental, and better disposed to yield to compression or extension by a slight alteration in their curvature.

The destructive irregularity of thick and thin works is partly averted by uncovering the thick parts of the casting, or even cooling them still more hastily, by throwing on water from watering-pots. In wheels this has been done by a hose, the axis of which is concentric with the wheel, the arms being all the time surrounded by the sand to retard their cooling; but it is the most judicious in all patterns, to make the substance for the metal as nearly uniform throughout as circumstances will admit, so as not to require these modes of partial treatment, which often compromise the ultimate strength of the casting.

Another mode sometimes adopted for avoiding the fracture of wheels, from the great dissimilarity of their proportions, is by inserting wrought-iron arms in the mould, but they do not always unite kindly with the iron of the rim and the nave. The same inconvenience occurs when iron pins are inserted in the ends of either of iron or brass castings, to serve for their attachment to their respective places: in iron castings it frequently produces the effect of chill casting, so as to render the works difficult to be turned or filed at the junction, and there is risk of the casting becoming blown or unsound in either case. When the pins are heated before being placed in the mould, they become nearly cold before the metal can be poured, and they also endanger the presence of a little steam or vapour, which is detrimental; therefore they are more generally put in cold, notwithstanding the sudden check they then give to the fluid metal.

The patterns for iron castings, of large size, are necessarily very expensive, especially those for hollow cylinders and pans, many of which are so large, that it would be impossible to find solid pieces of wood from which the patterns could be made; either with sufficient strength for present use, or with the necessary permanence of form for a subsequent period, as they would be almost sure either to break, or to become distorted from the effects of unequal shrinking, as explained by the diagrams, 13, 14 15, p. 19. Such patterns, therefore, require to be made of a great many thin layers or rings of wood, each consisting of 6, 8, or 12 pieces, like the felloes of wheels, so that in all parts the grain may be nearly in the direction of a tangent.

* Mr. Dodd had to contend with the shrinking of the nave, which was the last to cool; the accidents therefore occurred from the tension, instead of the compression of the spokes; this equally fatal effect was completely remedied by placing the arms as tangonts. - Trans. Soc. of Arts, Vol. LI., p. 66.

As they are glued up, every succeeding layer is connected with the former by glue and wooden pins or dowels, and the whole is afterwards turned to the tubular or hemispherical form, as the case may be. As the castings are generally required to be rather thin, such models are not only very expensive, but also very liable to accident; and besides it frequently occurs that only one or two castings of a kind may be required, which makes the proportional cost of the patterns excessive.

It fortunately happens, however, that this case, which is one of the most costly and uncertain by the employment of ordinary wood or metal patterns, becomes exceedingly manageable by a peculiar and simple application of the art of turning, (the one great center of the constructive arts, to which these pages are intended immediately and collaterally to apply:) and by which process, or one branch of loam moulding, to be explained in the following section, patterns are not generally required.