entirely on one side of the fulcrum knife-edge, or may extend from a point on one side of that knife-edge to some point on the other side of it-all without disturbing the equation. And, finally, the equation still holds for an unloading weighbeam-that is, one where the zero position of the poise is at that end of its travel which is farther from the weighbeam fulcrum, and where the weight indications represent load removed, rather than load applied as in conventional arrangements.

Expressed in words, our equation tells us that the weight of our poise must increase exactly in proportion as the weighbeam or poise capacity is increased, or as the fulcrum distance is increased, or as the multiple at the knife-edge of the load pivot is decreased, or as the run of the poise is decreased. Different people have different ways of remembering this relation; a weighbeam designer recently commented to me that he liked to remember "poise weight" times "poise run" times "multiple at load knife-edge" equals "weighbeam capacity" times "fulcrum distance."

A moment ago, passing mention was made of fractional poises, and these and the circumstances which necessitate them should be briefly considered. Especially for large-capacity scales of the beam type, it is rather common to furnish full-capacity weighbeams-that is, weighbeams on whose bars the entire nominal capacity of the scale is represented and with which no counterpoise weights are intended. to be used. As is the case with any other weighbeam, the value of the minimal graduation will be dictated by custom, trade demands, or purchase specifications. However, the necessity of representing the entire nominal capacity of the scale on the weighbeam itself, in conjunction with the requirements of "clear interval between graduations," readability, precision of positioning the poise at individual graduations, economical and/or convenient length of weighbeam, and other considerations, have led to two solutions, slightly different in appearance but fundamentally alike. These two solutions have in common the provision of a relatively heavy poise, known as a main poise, cooperating with a series of main graduations, and a relatively light poise, known as a fractional poise, cooperating with a series of subordinate graduations marked on a separate bar and, further, the requirement that the indications of both poises must be added to arrive at the weight value indicated. The essential difference between the two solutions lies in the fact that, in the one (see fig. 46) the fractional poise and the bar along which it is movable are parts of the main poise, while, in the other (see fig. 47) the two poises are not mechanically related, although both bars are necessarily integral parts of the weighbeam assembly.

The bar along which the fractional poise is movable may be either notched or smooth (sometimes the notches are formed in the fractional poise itself), but the bar along which the main poise is movable will practically without exception be notched. This is done in order to attain for the main poise a convenience of use and a ready and accurate means of positioning it at any normal weighing position which will be consistent with the precision of positioning the much lighter fractional poise and reading its weight indications. When notches are provided for any poise, of course, the seating of the pawl or latch into the notch, rather than the coincidence of index and graduation, governs

the position of the poise for any individual weight indication. For this reason, on a notched weighbeam bar, it is not strictly necessary that there be perfect coincidence of index and graduation when the poise is at any normal weighing position; however, in no case should the lack of absolute coincidence be such as to cause any question as to the indicated weight value.

For notched weighbeams of large-capacity scales, it is customary and proper to make the main notches at 1,000-pound intervals or 100-pound intervals and then to make the fractional-poise capacity equal to that interval-or, sometimes, to that interval minus one graduation on the fractional poise or bar. The latter is always done when the weighbeam is to be provided with type-registering means, due to practical con

struction requirements. Occasionally, what is known as a double-slide poise will be found, embodying two fractional poises; for example, if a small value is required for the minimal graduation and a relatively large value is assigned to the notch increment for the main poise, an unduly long fractional poise might result if this were to be single; in this case, the designer may elect to provide an intermediate fractional poise, the graduation interval of which will equal the total capacity of the smaller fractional bar, and the total capacity of which will equal the notch interval for the main poise.

On notched weighbeams, the distance center to center of notches on the same bar or poise will always be found to be some simple fraction of an inch (that is, for example, 1/2 inch rather than 1532 inch or 132 inch), and this is so on account of the characteristics of the machines used in this country for forming the notches. On main bars of large capacity scales, no closer spacing than six notches to the inch will ordinarily be found-this spacing being standard on weighbeams of railway track scales of modern design. Vehicle-scale weighbeams may be found with two, two and one-half, three, four, or five notches to the inch. On some small-capacity scales with notched weighbeams and single poises, a notch spacing of eight to the inch has been observed.

It may be noted in connection with notch-value and notch-spacing that a notched weighbeam will not and can not be made to indicate weight values in two different systems of weight units, as, for example, in both pounds or multiples thereof and kilograms or multiples thereof.

It is important that notch contour be uniform and that the pawl or latch tip fit snugly into each notch. After all, notches and pawls are but means to the end that successive normal poise positions be equidistiant within close limits. Most poises on notched weighbeams are provided with pawls which engage but one notch at a time, although a few poises will be found with what are known as multiplenotch pawls. As now being marketed, it is understood that the latter type of pawl will engage simultaneously some 15 separate notches, and it is obvious that, as compared with a poise having a single-notch pawl, the position of a poise provided with a multiple-notch pawl will be much less affected by wear, distortion, or imperfect location of any one individual notch.

And, finally, in figure 45, we note pivot C, known as the tip pivot or the counterpoise-hanger pivot. This pivot will not necessarily be found in weighbeams of the full-capacity type, but will be found in all weighbeams when all or part of the nominal capacity of the scale is represented by counterpoise weights. In any event, when this pivot is provided, the total multiple to its knife-edge should be a simple, definite, and known or readily determinable figure. Multiples such as 50, 100, 500, or 1,000 will be found to predominate, and in general are to be preferred. However, a few scales are understood to be in service where the multiple to the knife-edge of this pivot is 7,000 (or one of its larger factors)—the advantage of this figure being that the number of grains applied on the counterpoise hanger will balance exactly the same number of pounds applied on the load-receiving element, with resulting convenience in certain test routines, especially if the inspector has a set of grain weights but no decimal weights.

The counterpoise weights belonging to the scale and intended to be applied as necessary to the knife-edge of this pivot by means of the counterpoise hanger-and sometimes the hanger itself-are in the same category as the poise of which we have already spoken, in that they are mass standards, and that the scale itself is merely the means of comparison of the weight of the applied load with their weight. The scale may tell us-truthfully, withal-that a certain applied load weighs exactly 1,000 times as much as a certain counterpoise weightbut that does not particularly enlighten us as to the actual weight of the applied load unless we first know two things about the counterpoise weight (1) that it is accurate, and (2) that it is being used on a scale for which it was designed. Counterpoise weights, like poises, are essentially standards for comparison, and should be duly and accordingly respected.

Now, just a few observations as to the weighbeam in general. For reasons outlined earlier in our discussion, the knife-edges of the sevcral pivots will not necessarily all lie in the same plane; in fact, if they did, the weighbeam probably would not function satisfactorily. However, it is absolutely essential that each knife-edge be straight, and that all the knife-edges be parallel under any normal condition of loading—and, just as a reminder, this includes parallelism both vertically and horizontally. Pivots must be securely fixed in their respective positions in the weighbeam. The knife-edge line of the fulcrum bearing must be level, so that in use, the weighbeam will not tend to “walk” sidewise. There should be continuous contact between knife-edges and bearing surfaces for the full designed length, and for this a really selfalining type of bearing is essential. The bearing steels should be smooth, especially where they are intended to contact the knife-edges, and should be so formed as not to "pinch" the knife-edges. There should be no contact except along knife-edge lines between the weighbeam and any stand, loop, or bearing steel, or any interfering object. The beam rod should pull vertically against the knife-edge of the load pivot, and the beam rod, together with the shelf lever and the steelyard rod (when a shelf lever is used), must be clear of all interference. The contacts between loops and any hooks which may engage them should be flexible. No magnetism should be present in either the weighbeam or the trig loop. And the beam shelf should not deflect under load to an extent that will affect the weighing performance.

Obviously, there are many things about weighbeams which deserve careful attention, but which are outside of the scope of this presentation as announced in its introductory paragraph. If the interest of the members of the Conference warrants, it is to be hoped that this introduction to the subject may be followed up by treatment at subsequent Conferences of matters not discussed here. "Maintenance," "testing and adjusting," and "use" are three general, untouched subjects which suggest themselves at this time.

Even in this preliminary discussion, it has been necessary to cover a relatively large field in an extremely limited time, and many interesting and important matters have merely been mentioned in passing. To those in search of further written information on the subject or specific phases of it, I can commend two books, neither of which, however, is readily available in this country, especially under present conditions. I refer to H. C. Walters' excellent translation of Dr. Brauer's

"The Construction of the Balance" (published by the Incorporated Society of Inspectors of Weights and Measures, London), and to "A Treatise on Weighing Machines" by the late George A. Owen (published by Messrs. Charles Griffin & Co., Ltd., London). Mr. Owen's method of presenting some of the fundamentals of weighbeam design has been followed in this paper. In the reports of this Conference will be found much interesting and informative material, and particular reference is here made to "Notes on Scales" by C. A. Briggs (appearing on pages 135-175 of the report of the Eighth National Conference, held in 1913) and to "A Demonstration Showing the Necessity of Maintaining Scale Levers Level and the Lever Connections Plumb" by Charles C. Neale (appearing on pages 34-39 of the report of the Ninth National Conference, held in 1914).

LOCAL "MASTER" VEHICLE SCALES

PAPER OF LOUIS E. WITT, SEALER OF WEIGHTS AND MEASURES, CITY OF MILWAUKEE, WISCONSIN

The value of a local master vehicle scale is, of course, patently manifest to all of you as experienced weights and measures officials. Its applications to augment good weights and measures supervision are many and varied. It is the purpose of this paper to cite to you a few concrete examples of how such a master scale may be used to promote accuracy in large-capacity weighing and also how it may be used to safeguard the interests of the local or city government itself.

The Weights and Measures Department in the city of Milwaukee has had in operation a master motor-truck scale for more than 20 years. The scale is a 20-ton single-beam scale. It is in charge of an official in the Bureau of Weights and Measures who is known as the City Weighmaster. The Weighmaster is bonded and issues official city weight certificates, in duplicate, for each load weighed for a small fee. (We do not have in effect in Milwaukee a fee system for our regular weights and measures inspection service.) This master scale is maintained at a high degree of accuracy and efficiency. It is truly a master scale.

We in Milwaukee do not have a large-capacity vehicle-scale testing equiment of the type owned and operated by the National Bureau of Standards and which has recently been acquired by some of the State jurisdictions. Our heavy-duty testing equipment consists of a 2-ton truck with 5,000 pounds of 50-pound test weights. Our vehicle-scale tests, as a result, are made by sheer man power. This work of testing all the commercial heavy-duty scales and vehicle scales in the city takes a little more than 2 months each year, using a crew of two weights and measures inspectors and two temporary laborers who have been hired for that purpose. The inspectors assigned to this duty give each vehicle scale in the city a very careful and thorough inspection, and test most of these scales almost to capacity by the use of heavy strain loads which are usually available.

This brings us to the first use to which our scale is put as a "master scale." Each morning, before starting out on the work of testing vehicle scales, the inspector in charge of the work obtains the gross weight of our vehicle-scale testing outfit by weighing it on the master scale. It is obvious that this gross weight must be accurate because it is used as a test weight.