percents of underweight items in the lot (or lots) how the odds change for getting a run of ten samples all averaging underweight. The odds are 1,023 to 1 against such a run if the lot is O.K. The odds are one to one if about two-thirds of all the items are underweight. Persistent underweight packing quickly reveals itself in the sample averages.

Note that there is no need to get individual weights. There is no bother about tolerances. There is only an insistence that the item weights be symmetrically distributed about the correct weight. This will happen if the packing is performed so that the lot average is held close to the correct value. Modern quality control methods are well established to make it easy to meet this requirement. Attention is now focused, not on the individual item, but on the overall performance.

The same approach may be applied to the examination of individual items. Table 2 shows how the chance of getting a run of items all underweight depends on the percent of underweight items in the lot being sampled.

TABLE 2. Odds against getting a run of items all underweight

This table reveals that the odds are very much against getting ten (or seven) successive items underweight if the lot is at all close to having equal proportions of over and underweight items. Long runs of underweight items are not likely to turn up unless three-fourths or more of the items are underweight. If there is this percent of underweight items, it implies an average content below the correct value. All that the inspector is doing is to apply a well-known statistical technique that could equally well be used by the packer to keep his average where it belongs. We expect a fair coin to give equal chances for heads and tails. A run of heads arouses suspicion. Ultimately the player makes a choice. Either the coin is fair and a very improbable sequence of events has occurred, or else the game is rigged. A long run of either heads or tails leads to the latter conclusion, especially if playing with a stranger.

Summary

This discussion has made the following points:

(a) That the use of samples is an effective method of checking packaged items.

(b) That individuals experienced in statistical quality control should be consulted for the interpretation of the data obtained from samples.

(c) That agreements should be reached regarding whatever action should be taken on the basis of the sample results.

(d) That the need for setting tolerances may be avoided by substituting the simple requirement that the contents of the items packed should be symmetrically distributed around the correct value.

REPORT OF COMMITTEE ON SPECIFICATIONS AND TOLERANCES

Presented by T. C. HARRIS, Jr., Supervisor, Weights and Measures Section, Department of Agriculture and Immigration, State of Virginia

The Committee on Specifications and Tolerances of the 48th National Conference on Weights and Measures submits its report-a tentative report as amended by this final report.

1. GENERAL CODE

1.1. G-T.1. Acceptance tolerances.-It has been suggested to the Committee that the provisions of general tolerance paragraph G-T.1. might be so revised as to make subsections (c) and (d) somewhat more practical and enforceable. It has been pointed out that the "3 months" period specified in (c) and (d) for overhauled or repaired equipment, appears to be somewhat excessive when compared with the new equipment usually anticipated under (a) and (b). The Committee believes that this suggestion has considerable merit and will propose the reduction of the 3 months period in (c) and (d) to "2 months” and "1 month" respectively.

Another suggestion reaching the Committee relative to G-T.1. has to do with the removal of a permanently installed device from one fixed location and its reinstallation in another fixed location in the same State. It was suggested that devices in this category be included under subsection (c). The Committee concurs in this suggestion.

To implement the revisions explained above, the Committe recommends that general tolerance paragraph G-T.1. of the General Code be amended to read as follows:

G-T.1. ACCEPTANCE TOLERANCES.-Acceptance tolerances shall apply to equipment of the following classes:

(a) Equipment that is about to be put into use for the first time in the State.

(b) Equipment that has been put into use within the preceding 3 months and is being officially tested for the first time.

(c) Equipment that is being officially tested for the first time within 2 months after major reconditioning or overhaul, or reinstallation in a new fixed location.

(d) Equipment that is being officially tested for the first time within 1 month after repair, adjustment, or other corrective service operation, following official rejection.

(Item 1, General Code, was adopted by voice vote.)

2. SCALE CODE

Since the 47th National Conference in 1962, the Committee has diligently studied all recommendations and suggestions received from

weights and measures officials, including State and regional weights and measures associations, the Scale Manufacturing Association, and others relative to the Scale Code. In its tentative report the Committee has attempted to implement these suggestions, especially those pertaining to simplification, clarification, and adjustment of tolerances. However, in light of the testimony presented during the open hearing on June 10, the Committee is convinced that this item should be given further study. Therefore, the Committee now recommends that item 2 be continued on the agenda of the Committee for another year.

(Item 2, Scale Code, was adopted by voice vote.)

3.1. SLOW-FLOW METERS

3. LIQUID-MEASURING DEVICE CODE

3.1.1. General.-In recent months there has been an increasing amount of interest in the so-called "slow-flow" meters, intended for installation in individual homes or at individual mobile homes in a trailer camp, for the commercial measurement, cumulatively, of the volume of liquid fuel being consumed. The fuel is supplied from a central storage facility. The rates of flow through such meters may range up to a maximum of 10 gallons per hour.

Normally, these devices would not be covered by public utility laws, but would fall within the jurisdiction of weights and measures statutes. For this reason, "slow-flow" meters are of interest and concern to weights and measures officials.

3.1.2. Tests and equipment. The National Bureau of Standards has conducted extensive studies and tests of these new devices, in recent months. It has been concluded that special testing equipment is essential in the conduct of these tests. The Bureau has developed such special testing equipment and has conducted extensive tests on meters of four different manufacturers. The tests were designed to (1) prove the effectiveness of each of several special testing systems, (2) develop test procedures for slow-flow meters, (3) determine the accuracy of the meters, and (4) gather data necessary in the development of technical requirements for the meters.

It is expected that a detailed description of proper testing equipment and specific test procedures will be available from the Bureau's Office of Weights and Measures in the near future.

3.1.3. Nature of tests. Owing to the intricate nature of these tests, connected with the fact that an installation of slow-flow meters may include as many as several hundred individual devices, tests are contemplated as laboratory operations, and testing undoubtedly will be accomplished on a sample basis. The individual testing of these devices in service appears to be completely impractical and uneconomical.

3.1.4. Code amendments.-It is evident that a few amendments to the Liquid-Measuring Device Code will be necessary to make it clearly adequate in its coverage of slow-flow meters. The major amendments, however, may be expected to include (1) special indicating facilities for test purposes, (2) special provisions concerning test procedures, (3) special marking requirements, and (4) proper tolerances to be applied to these new and different devices. On the other hand, many of the present requirements of the General Code and the Liquid-Measur

ing Device Code will be applicable and should be enforced in all instances where they are clearly applicable. The Bureau's Office of Weights and Measures, based upon its tests and experiences thus far, has developed a tentative table of tolerances for slow-flow meters. This table is included in this report.

It is also suggested by the Committee that this table of tolerances will be useful to weights and measures officials and users of slow-flow meters, to serve as a guide during the interim awaiting final adoption by the Conference of all necessary amendments in this area.

The Committee Chairman made the following motion relating to item 3.1:

(1) That item 3.1. of the tentative report be adopted; (2) that paragraph 3.1.5, which is the table of tolerances in the Committee's tentative report, be adopted as the tentative table of tolerances for slow-flow meters; and (3) that the National Bureau of Standards be requested to print and distribute in looseleaf form the appropriate information contained in the paragraphs 3.1.1, 3.1.2, 3.1.3, and 3.1.4 of item 3.1 of the tentative report.

The motion was duly seconded.

(After considerable floor discussion and explanation by the Committee Chairman, item 3.1., slow-flow meters, was adopted by voice vote.)

3.2. MILK METERS

3.2.1. General.-Practically since the first use of farm milk tanks, there has been considerable interest in the dairy industry concerning the possibility of using meters, that would meet all weights and measures and sanitation requirements, for the measurement of fluid dairy products. This interest has continued to increase with the rather rapid expansion in the use of farm milk tanks and the bulk handling of fresh whole milk. Although the Committee is not aware of the availability of any specific complete metering system that has been clearly demonstrated to be entirely satisfactory for this purpose, it does appear inevitable that one or more such systems will be developed and made available in the not too distant future. The Committee does understand that considerable study and tests have been

conducted at Pennsylvania State University, University Park, Pennsylvania. Also, the National Bureau of Standards reported to the 47th National Conference relative to its experiences in this field.

3.2.2. Tolerances.-It now appears reasonably certain that these special meters for the measurement of milk can be expected to perform with the accuracy requirements now specified for wholesale liquid-measuring devices and vehicle tank meters. Also, it appears evident that many of the requirements contained in the General Code, the Liquid-Measuring Device Code, and the Vehicle Tank Code will serve adequately in the future official control of milk meters. These requirements should be considered and applied by weights and measures officers, insofar as the requirements are clearly applicable to milk

meters.

3.2.3. Commercial use.-It would be anticipated that a large majority of commercial milk meters would be installed upon the vehicle tanks used in the collection of milk from the individual dairy farms. In some instances, however, it is probable that commercial use will be made of these meters permanently installed in a milk plant.

3.2.4. Special problems. One of the complications offered by a meter for this use is a result of the very strict requirements of sanitation officials. These sanitation requirements make it necessary that a meter be dissassembled for cleaning and inspection. This situation dictates that all individual parts associated with a given meter should be permanently marked for indentification purposes.

Another problem peculiar to the operation of milk meters on vehicle tanks is that of the proper priming of the system prior to the first commercial measurement of product after the device has been disassembled and cleaned. Some people have expressed the opinion that the system should be properly primed before use, regardless of any inconvenience that might be involved. On the other hand, others have expressed the possibility of advanced determination, by the weights and measures official, of the actual volume involved in the proper priming of the device. In this case, the dairymen involved would be allowed, over and above the quantity registered by the meter, the amount required in the priming process. In the opinion of the Committee, this is a very important point and should be given careful consideration by all interested weights and measures officials and members of the dairy industry. Comments, suggestions, and recommendations in this area will be appreciated.

3.2.5. Testing fluid.-The Committee is not aware of the availabilty of any substitute for milk to be used as the official testing fluid. It is possible that some substitute material may be developed. At the present time, however, it appears that it would be necessary to use natural whole milk for official testing purposes.

3.2.6. Standards and equipment.—In the official testing of milk meters, it is evident that a weights and measures department would be required to obtain special sanitary standards and equipment. It is expected that the National Bureau of Standards will be able to furnish consultation and advice in this area through its Office of Weights and Measures, by the time such metering systems are employed in commercial use.

3.2.7. Code amendments. It is evident that certain amendments to the Liquid-Measuring Device Code and the Vehicle Tank Code will be necessary. It is the expectation of the Committee to continue to