(4) The values for Mi, R, R, Mi, X, and (c) X and R control charts. Control charts consist of an X chart and may consist of an R chart. When both are used they are contained on the same sheet with the limits for the X chart on the top portion and the limits for the R chart on the lower portion with a space separating the two such that plottings from one does not overlap those of the other. Separate control charts shall be maintained for each item. More than one day's production of an item may be plotted on a single control chart. (1) charts. (i) The X charts for two-sided specifications shall consist of lines and values properly representing both the upper and lower reject and warning limit(s), and may require the use of minimum and maximum specification lot averages. (ii) The X chart for one-sided specifications shall consist of lines and values properly representing, as applicable, the upper or lower reject and warning limit(s), and the minimum or maximum specification lot average. (2) R charts. When the R chart is used, it shall consist of lines and values properly representing R', Rmax, and the lower limit for R. For subgroup sizes of 6 or less the value of the lower limit for R will always be zero. [38 FR 10447, Apr. 27, 1973; 38 FR 12729, May 15, 1973] PROCEDURE § 52.206 Sampling. (a) General. In order to obtain the most reliable results, sampling should be performed at a point where no further change in the variable can occur. Sample units or subgroups shall be drawn separately for each code, container size, and style. (b) Subgroup sampling. Sampling by subgroups consists of drawing more than one sample unit at approximately the same time. This type of sampling is required in the use of the conventional X and R determinations for variables. When a single code is processed on more than one line sumultaneously, it is desirable to obtain a subgroup representing a single line, alternating the lines each time a subgroup is drawn for that code. If the processing procedure prohibits obtaining a subgroup representing a single line, it shall be necessary to sample by code only, omitting line identity. Each subgroup thus obtained will generally represent more than one line and consequently reflect variations in the code as a whole but not variations for a single line. (c) Sample unit sampling. This procedure consists of drawing a single sample unit at random at specified intervals from a production. This type of sampling may employ the use of a moving average and is applicable to a variable or a process when subgroup sampling is not feasible. (a) General. Immediately after the sample unit or subgroup has been taken, measurements and calculations shall be made and recorded on the appropriate form as required. (b) Recording data on the X and R data sheet. The data to be recorded on the X and R data sheet is as follows: (1) The time the sample unit or subgroup is taken; (2) The value for each individual measurement (X value); (3) The total value for all the X values in the subgroup, when applicable; (4) The average value (X) for each subgroup, when applicable; (5) The Mi value for each subgroup, when applicable; (6) The range value (R) for each subgroup, when applicable; (7) The sample median value (MI) when the median is used; (8) The sample average value (X) when required; and (9) The average range value (R) when applicable. (c) Recording data on the X and R control chart. The data to be recorded on the X and R control chart is as follows: (1) Individual measurements (X values). The values representing the individual measurements of each subgroup (X value) shall be plotted on the X chart. The distance between the smallest and largest value of the individual measurements in the subgroup may be used for R in lieu of plotting the range on the R chart. It is desirable to plot all the values for a subgroup on the same linear segment of thec hart. Identical values are plotted adjacently in such a manner as to associate such values with the subgroup in which they occur. (2) The median. When the median option is used, the median for each subgroup shall be identified on the X chart in such a manner as to make it readily distinguishable from the rest of the value in the subgroup. When a subgroup consists of an even number of individual measurements, the median represented by the arithmetic mean of the two middle values shall also be plotted on the X chart in such a manner as to associate it with the subgroup from which it was calculated. When the median option is used for a particular production, it is not permissible to change to the option for conventional averages during the same shift. (3) Subgroup averages (X values). When the option for conventional averages is used, the average of each subgroup is calculated and recorded on the X and R data sheet. The subgroup average is then plotted on the chart in such a manner as to associate it with the subgroup from which it was calculated. When this option is used, it is not permissible to change to the median option during the same shift. (4) The range (R value). The range of each subgroup, when required, shall be plotted on the R chart. (5) The average median Mi value) and sample average (value). When the plotted values obviously indicate the acceptability of the lot, Mi or X need not be obtained. When the plotted values for Mi or X or both indicate the acceptability of a lot may be questionalbe, Mi in the case of the median option, or in = the case of the option for conventional averages must be obtained and recorded. Acceptability of the lot with respect to the sample median or sample average may then be determined in accordance with acceptance criteria as prescribed under § 52.209(a) of these standards. [38 FR 10447, Apr. 27, 1973; 38 FR 12729, May 15, 1973] § 52.208 Application of sampling allowance. The amount of sampling allowance is obtained from the sampling allowance chart in these standards. The actual value for the sampling allowance is obtained from the point of intersection of a horizontal line drawn from a point on the vertical axis representing the sample size to the diagonal line for the item involved. The value on the horizontal axis corresponding to the vertical line nearest to this point of intersection is the sampling allowance. The sampling allowances are applied to the specified lot averages to allow for deviations from these averages that are inherent in any sampling procedure. These allowances are applied to the specified lot averages only when the sample. average or Mi) is less than 'miz or greater than 'max. When this occurs, the specification average is adjusted by subtracting the allowance from X'min or adding to X'max, whichever is applicable. § 52.209 Acceptance and rejection criteria. (a) Conventional averages and the median. (1) Acceptance. A lot shall be accepted as meeting specified requirements for a variable provided that: (i) All X values are equal to or greater than LRL and equal to or less than URL; (ii) All Mi or X values are equal to or greater than LRL and equal to or less than URLY; (iii) Mi or X is equal to or less than X'max adjusted and equal to or greater than X'min adjusted; and (iv) The condition under paragraph (a) (2) (ii) of this section does not exist. (2) Rejection. A lot shall be rejected for failure to meet specified requirements for a variable under the following conditions: (1) Failure to meet any of the requirements under paragraph (a) (1), (i), (ii), or (iii) of this section; or (ii) All Mi or X values are less than X'min or greater than 'mai. (b) Moving averages. (1) Acceptance. A lot shall be accepted as meeting specified requirements for a variable provided that: (1) All X values are equal to or greater than LRL and equal to or less than URL; (ii) All X values are equal to or greater than LRL and equal to or less than URLX; (iii) X is equal to or less than 'max adjusted and equal to or greater than X'min adjusted; and (iv) The condition under paragraph (b) (2) (ii) of this section does not exist. (2) Rejection. A lot shall be rejected for failure to meet specified requirements for a variable under the following conditions: (i) Failure to meet any of the conditions under paragraph (b) (1) (1), (H), and (iii) of this section; or (ii) All X values are less than X'mi or greater than 'max. [38 FR 10447, Apr. 27, 1973; 38 FR 12729, May 15, 1973] SAMPLING ALLOWANCE CHART § 52.210 Sampling allowance chart. (a) The variables inspection standards may be used, where applicable, by any food processor for quality control purposes. Certification by the U.S. Department of Agriculture (USDA) of fill weights based on these standards, however, is applicable only under on-line inspection when a USDA inspector is present at time of packing and maintains appropriate records of the filling process. (b) The procedure for determination and certification of fill of container, based on drained weights, may be used in the case of lot inspection as well as in-line inspection. This procedure, where applicable, is outlined in the U.S. standards for the respective products. § 52.223 Options. The options of the median or conventional average may be used for determining compliance of fill weights with a specified minimum average fill weight. § 52.224 Type of specification and applicable limits. (a) For the determination of fill weights, a low-sided specification applies. (b) The specification limits, and values for such limits, are specified in those U.S. Standards for Grades of canned fruits, vegetables, and related products which incorporate the fill weight procedure. Suggested specification limits, and the values for such limits, for those products for which fill weights have been established-but are not yet incorporated in the USDA grade standardsmay be obtained as supplements to these standards upon request to: Chief, Processed Products Standardization and Inspection Branch, Fruit and Vegetable Division, AMS, U.S. Department of Agriculture, Washington, D.C. 20250. (c) All of these limits, whether stated in the USDA grade standards or supplements hereof, are to be applied in accordance with the proposed U.S. Standards for Inspection by Variables. § 52.225 PROCEDURE Preparation of forms. (a) Prepare the data sheet and control charts prior to the start of a processing period. Draw lines representing the X'min, LRL, and LRL values applicable for the product, container size, and style on the appropriate linear segment of the X chart in such a maner as to be clearly visible when posted in the plant. Lines representing LWL and LWL may be drawn on the chart at the option of the user; however, all the values representing each limit (Xmin, LWLx, LRLx, LWL and LRL) are placed on the X chart adjacent to the appropriate limit and identified (in brief form) as X', WL, RLX, WL, or RL accordingly). (b) When the R chart is used, it is desirable to place it on the lower portion of the control chart, using the bottom as zero, which is the lower limit for R. (c) Space the limits for R' and Rmax on the appropriate linear segment above the zero line. The upper portion of the control chart is then used for the X chart. (d) Attachment 1 is an example of an X and R data sheet. Attachment 2 is an example of a control chart which combines the X chart and the R chart. However, use of the R chart is not required. The chart illustrates the use of the median option as prescribed in the proposed U.S. Standards for Inspection by Variables. The dots plotted on the X chart represent the X values for the subgroup. The encircled dot represents the median of each subgroup. For easier identification the circle should be in a different color than the dots. [38 FR 10450, Apr. 27, 1973; 38 FR 12729, May 15, 1973] § 52.226 Posting control charts. It is desirable to post the control charts in such a place in the plant as to be clearly visible to the filler operator as (a) General. Variations in the weight between individual empty containers (tare) must be taken into consideration when determining fill weights. These variations may be accounted for by procedure 1 or procedure 2 of this section. (b) Procedure 1. Determine tare weights of each container size at least twice a week. In addition, determine tare weights whenever it is suspected that the present tare weight is no longer applicable. Some of the more common reasons for checking tare weights are changes in supply of containers, thickness of tin plate, and type of container. (1) The number of empty containers taken at random from a given lot of containers to determine the average tare weight is as follows: For containers equivalent to: No. 21⁄2 container and smaller than 10 empty containers. Larger than No. 21⁄2 than 15 empty containers. not less not less (a) Subgroup size. Unless otherwise specified, the subgroup size shall be five sample units. (1) The control chart values, except for X'min, will always be based on the specified subgroup size. When a subgroup size other than that specified is used, the values for the applicable limits must be changed accordingly. (b) Time and point of sampling. The subgroups are drawn at a point in the process where no further change in the fill weight of the fruit or vegetable ingredient can occur. (c) Sampling frequency. Unless otherwise specified, the sampling fre quency for all canned fruits, vegetables, and related products, for which fill weights have been established shall be approximately every 45 minutes. (1) The sampling frequency should be arranged to obtain subgroups at varying intervals and drawn in such a manner that the filler operators will not know when a subgroup will be drawn or which containers will be in the subgroup. This is important so that a nonbiased sample will be selected which will reflect the normal filling process. [38 FR 10450, Apr. 27, 1973; 38 FR 12729, May 15, 1973] (a) Immediately after the subgroup has been drawn, invert each container on the hand or sieve and allow the product to drain until free from liquid, but in no case less than 10 seconds. In this step, do not remove the product from the container when procedure 1 as prescribed in § 52.227 is used. The sample units are then weighed and the net weight of the fruit or vegetable ingredient (total weight minus tare weight) for each sample unit is recorded on the X and R data sheet. (b) In the case of procedure 1 if it is suspected that the weight of individual empty containers deviate from the average tare weight to cause an individual measurement to fall slightly below (or slightly above) LRL, the following procedure may be used: (1) After the X value has been obtained as previously described, empty the contents of the suspected container and wipe it clean and dry. (2) Weigh the suspected dry container. The amount of deviation is the difference between the weight of the individual container and the average tare weight. (3) Adjust the suspected X value by the amount of deviation thus found. (c) When adjustments for tare weight are made for individual measurements that fall slightly below LRL, adjustments must also be made for individual measurements that are slightly above LRL. § 52.230 Calculations and recording data. (a) Immediately after the measurements have been obtained, the required |