face is quick to assume a more natural position as soon as it is heated.

All tapped or hobbed dies have practically four errors. First, the lead screw error of the engine lathe in which the tap or hob is made. Second, the error of hardening the tap or hob. Third and fourth, the double error of changing form of the die in hardening it, due to releasing the compressed metal, and the usual hardening change that takes place at the same time.

By the use of the present method, all errors existing in the milling cutter are corrected in the milling machine, hence the lead is only affected by the final hardening of the chaser, which takes place under such favorable conditions that it produces no appreciable effect.

The outcome of this scheme has not only been a more accurate production of the screw-thread but it has also made it practicable to measure a screw-thread by a reliable means.

All that has been said of the impossibility of making correct leading dies by other methods is equally true of making the so-called thread gauges, consisting of a piece of steel tapped and hardened, but sometimes furnished with a means intended for adjustment for wear but in reality of no value, for it is impossible to adjust such gauges in the direction of their wearing. A more deceptive and unsatisfactory gauge could hardly be devised. Every workman knows that one piece of work may fit a gauge of this kind loosely and yet not enter the tapped hole, and another piece of work may fit it closely and yet rattle in the same tapped holes owing to difference in lead and shape.

The three dimensions of a screw-thread, its shape, lead, and diameter, should be measured separately, if the object of such measurement is to detect the error for the purpose of correcting it. The lead can be measured by placing an accurate scale on a screw-thread of from three to eighteen inches in length, according to accuracy required. The shape is best measured by an ordinary template; after the lead and shape are known to be practically correct, the diameter may be obtained by a flatpointed micrometer gauge, ring or snap gauge, at the top of the thread.

In practice, the place and time to correct the lead and shape is in the die when it is made, and then the diameter may be readily gauged as the screws are made, it being necessary to occasionally take a test of the lead as the die becomes worn.

DCCLIV.*

A STAY BOLT THREADING DEVICE.

BY JAMES HARTNESS, SPRINGFIELD, vt.

(Member of the Society.)

THIS paper describes a means for threading stay bolts that was mentioned by the writer in the discussion of a paper entitled "Experiments in Boiler Bracing," presented by Mr. Francis J. Cole at the last meeting.

The scheme may be briefly and perhaps completely described as tandem dies for simultaneously threading both ends of a stay bolt to insure an accurate correspondence in lead; the details of which may be mentioned as a means for accurately adjusting the relative longitudinal position of the dies, and employment, at least for the forward die, of some type of opening die not too great in length and not too inaccurate in lead, and possessing good diameter controlling features.

The drawing (Fig. 26) shows the use of the automatic die described in a paper by the writer, entitled "A Screw Die for the Turret Lathe."

Both dies employed should be of the opening type, but the scheme would be equally as accurate if a solid or non-opening die were to be used in the place of the rear die. In fact, if the time consumed in operation were not to be taken into consideration, the front die could be non-opening and used to cut both ends, and in the place of the rear die a nut suitably mounted could be used.

The plate to which the dies are affixed is not necessary when the scheme is used in the flat turret lathe, for in that machine the dies may be attached directly to the turret without the use of the plate.

A brief consideration of the other methods and their results may assist in setting forth the value of this tandem die.

Stay bolts are mostly cut by dies permitted to control their own

* Presented at the New York meeting (December, 1897) of the American Society of Mechanical Engineers, and forming part of Volume XIX. of the Transactions.

lead. Occasionally, however, a lead screw is used to govern the pitch of the die. The process of making the dies that have been used for stay bolt threading is unreliable, as shown in the paper entitled, "A Screw Die for the Turret Lathe," and the lead screw scheme seldom controls the lead of the die at the beginning of its cut on account of the slackness of the slides and intermediate connections, hence the die usually has a chance to cut a very

The springing of the intermediate parts is mentioned because a die that has become a trifle dull will offer great resistance to any means employed to vary its lead.

The excessive clearance of a cutting die as led by lead screws makes the accurate maintenance of size difficult, and also gives a die the tendency to cut "out of round."

DCCLV.*

MACHINE MOULDING WITHOUT STRIPPING PLATES.

BY E. H. MUMFORD, PLAINFIELD, N. J.

(Member of the Society.)

MOULDING machines may be classed under three heads. First, machines which only ram the moulds, and, when the ramming is done by means of a side lever, by hand, are generally called "squeezers." Second, machines which only draw the patterns, the ramming being accomplished by the usual hand methods. Third, machines which both ram the moulds and draw the patterns, ramming either by a hand-pulled lever or by fluid pressure on piston or plunger, and drawing the patterns through a plate called a "stripping plate" or "drop plate "-till recently the usual method-or without the use of this plate fitting everywhere to pattern outline at the parting surface, the patterns being effectively machine guided in either case.

It is to the third class that the machine which is used to illustrate the subject of this paper belongs, and which would seem to have enough which is novel in the application of machinery to the foundry to merit the attention of the Society. At the risk of appearing pedantic, but with a view to developing an appreciation of the true function of the method of patterndrawing used in this machine, attention is called to the following sectional views of moulds and ways of drawing patterns occurring in machine moulding. Fig. 27 shows an ordinary "gate" of fitting patterns being drawn from the drag or nowel part of the mould by means of a spike and rapper wielded by the moulder's hand after cope and drag have been rammed together on a "squeezer" and the cope has been removed. Frequently the pernicious "swab " is used to soak and so strengthen joint outlines of the sand before drawing patterns, in such cases as this. In this case, before the cope is lifted, these patterns must be vigorously rapped through the cope; an amount depending *Presented at the New York meeting (December, 1897) of the American Society of Mechanical Engineers, and forming part of Volume XIX. of the Transactions.

(and so does the size of the casting) upon the mood and strength of the moulder.

Fig. 28 shows the stripping or drop plate method of drawing patterns.

In this method, the patterns are not rapped at all and are drawn in a practically straight line so that the mould is absolutely pattern size.

The stripping plate is fitted accurately to every outline at the joint surface of the patterns, obviously at considerable expense, and, of course, at the instant of drawing the patterns, supports the joint surface of the mould entirely. This is, at first sight, an ideal method of drawing patterns, and it has for years

been the only method practised on machines. It has two disadvantages. The patterns are separated from the stripping plate by the necessary joint fissure between the two. Fine sand continually falls into this and, adhering to the joint surfaces more or less, grinds the fissure wider. This leads to a gradual reduction of size of patterns on vertical surfaces and a widening of the joint fissure often to such an extent that wire edges are formed on the mould, causing, on fine work, "crushing" and consequently dirty joints. A nicely fitted but worn plate of twenty-four pieces, which had cost, at shop expense only, $250, was recently replaced by a plate of twenty-eight pieces, fitted ready for the machine under the new system about to be described, for not more than $25.

The stripping-plate method has another drawback, not always